Investigation of Fugitive Emissions from Petrochemical ... · PDF fileEPA/600/R-09/136...

EPA/600/R-09/136September 2009

Investigation of Fugitive Emissions from Petrochemical Transport Barges Using Optical Remote Sensing

Final Report

EPA/600/R-09/136September 2009


by

Eben Thoma Air Pollution Prevention and Control Division

National Risk Management Research Laboratory Research Triangle Park, NC 27711

National Risk Management Research Laboratory Office of Research and Development U.S. Environmental Protection Agency

Cincinnati, Ohio 45268


September 2009

TABLE OF CONTENTS List of Appendices ii

List of Tables iii

List of Figures iv

List of Acronyms v

Executive Summary vii

1. Introduction 1

1.1 Background 1

1.2 Project Description 1

2. Description of Test Sites, Measurement Methods and Site Deployment 4

2.1 Aerial PGIE Observations 5

2.2 Ground-based PGIE Observations 7

2.3 Scanning OP-FTIRs and OTM 10 Protocol 7

2.3.1 OTM 10 and the Vertical Radial Plume Mapping Method 10

2.3.2 OTM 10 Fugitive Emission Quantification 14

2.3.3 Supporting Measurements for Ground-based ORS 14

2.3.3.1 Meteorological Data 14

2.3.3.2 OP-FTIR Instrument-Mirror Distance 15

2.3.4 PIC Emission Measurements with OP-FTIR Instrument 15

2.3.5 Meteorological Data Collection with the R.M. Young Heads 16

2.3.6 Optical Path Length Determination with the Topcon Theodolite 17

2.3.7 Calculating Emission Flux using the VRPM Method 17

2.3.8 Site Deployment Description at the Lock 18

2.4 Bagging Tests 18

3. Aerial and Ground-based PGIE Results and Discussion 20

3.1 Aerial PGIE Observations by LSI 20

3.2 On board Barge PGIE Observations by LSI 24

3.3 PGIE Observations From the Lock Wall 34

4. OTM 10 AM Emission Flux and Trace Compound Speciation Results 42

4.1 Data Graphs and Tables for Select Events 42

4.1.1 Summary of the Results of Analysis of Trace VOC Concentrations 54

4.2 Instances of Emissions Detected with the PGIE but not with ORS Measurements 54

i

September 2009


4.3 Evaluation of AM Emissions from Tugs 55

5. Bagging Test Emission Estimate Results 57

6. Comparison of OP-FTIR and Bagging Test Emission Flux Results 58

7. Quality Assurance/Quality Control 61 7.1 Instrument Calibration 61

7.2 Assessment of DQI Goals 61

7.2.1 DQI Check for Analyte PIC (OP-FTIR) Measurement 62

7.2.2 PGIE Relative Opacity DQI Assessment 63

7.2.3 Meteorological Head DQI Assessment 63

7.2.4 Topcon Theodolite DQI Assessment 64

7.2.5 QC Checks of OP-FTIR Instrument Performance 64

7.3 Estimate of Uncertainty in the OTM 10 Emission Flux Measurements 65

7.4 Uncertainty in the LDEQ Leak Bagging Estimates 67

7.5 General Data Limitations 68

7.6 Deviations from the QAPP 68

8. Summary 69

9. References 72

List of Appendices

A. LSI Report: Leak Detection using LSI Infrared Gas Imaging, BEM 1 Barge Study; Aerial Imaging (October 27, 2008)

B. LSI Aerial PGIE Images

C. LSI Report: Leak Detection using LSI Infrared Gas Imaging, BEM 1 Barge Study; Ground Crew Survey (October 21, 2008)

D. LSI Ground Survey PGIE Images

E. LDEQ/ARCADIS Lock Wall PGIE Images

F. Alkane Mixture (AM) Measurement by OP-FTIR

G. OTM 10 Data Graphs and Tables

H. LDEQ Onboard Leak Bagging Test Report: Barge Emissions Measurement Project Final Report; SAGE Environmental Consulting (December 29, 2008)

I. Comparison of Carbon Monoxide and Alkane Mixture Concentrations for 9 Barge Emissions Events to Investigate the Contribution of Emissions from Tugs

J. Response to Comments from the American Waterways Operators

ii


September 2009

List of Tables

Table 2-1. Relative Response of Hydrocarbon with LSI Infrared Imaging Camera (FLIR) 6

Table 2-2. Target Compound List 16

Table 3-1. Summary Table of Barge Leaks Identified by LSI Aerial Survey 20

Table 3-2. Summary Table of Barge Leaks Identified by LSI Ground-based Observations 25

Table 3-3. Summary Table of Barge Leaks Identified by LSI Lock Wall Observations 35

Table 3-4. Summary Table of Barge Leaks Identified by LDEQ/ARCADIS Lock Wall Observations 38

Table 4-1. 9/24/ 2008 ─ Event #1 43

Table 4-2. AM Flux Values Measured during 9/24/ 2008, Event #1 43

Table 4-3. 9/29/ 2008 ─ Event #2 44


Table 4-5. 10/1/ 2008─ Event #2 46


Table 4-7. 10/2/ 2008 ─ Event #2 48


Table 4-9. 10/5/ 2008 ─ Event #1 50


Table 4-11. 10/9/2008- Event #7 52


Table 4-13. Summary of VOC Analysis 54

Table 4-14. Summary of Leak Events Detected by the PGIE but not ORS Instrumentation 55

Table 5-1. Summary of LDEQ Bagging Test Results 57

Table 6-1. Summary of LDEQ Bagging Test Barge Totals and Most Significant OTM 10 Results 59

Table 7-1. Instrumentation Calibration Frequency and Description 61

Table 7-2. Data Quality Indicator Goals for the Project 62

Table 7-3. Results of Flux Values Calculated by the VRPM Fit Explorer Program With a Fixed Peak Plume Concentration Location and Varying Values of the σy Parameter 66

Table 7-4. Results of Flux Values Calculated by the VRPM Fit Explorer Program with a Fixed σy Parameter and Varying Peak Plume Concentration Locations 66

iii


September 2009

List of Figures

Figure 2-1. U.S. Army Corps of Engineers Lock Study Site and Measurement Configurations 9

Figure 2-2. Images Illustrating OTM 10 Setup and Barges at the Port Allen Lock 10

Figure 2-3. General OTM 10 VRPM Measurement Configuration 11

Figure 2-4. Representation of Lock Cross Section and Wind Flow 13

Figure 2-5. IMACC OP-FTIR Instrument and Scanner 15

Figure 2-6. Images from LDEQ Leak Bagging Study 19

Figure 3-1. Example #1 from LSI Helicopter Survey – Leak from Vent on Aft Side of Barge A27 22

Figure 3-2. Example #2 from LSI Helicopter Survey – Leak from Vent Stack in Center of Barge A34 23

Figure 3-3. Example #3 from LSI Helicopter Survey – Leak from Top Hatches on Bow of Barge A31 23

Figure 3-4. Example from LSI Ground Survey – Sampling During Bagging Test 28

Figure 3-5. Leak from Cargo Hatch on Barge G2- Mass Leak 1.86 g/s 28

Figure 3-6. Leak from Ullage Hatch on Barge G4- Mass Leak 0.31 g/s 29





Figure 3-11. Leak from Vent on Barge G6- Mass Leak 1.45 g/s 31

Figure 3-12. Leak from Cofferdam Hatch on Barge G6- Mass Leak 1.99 g/s 32

Figure 3-13. Leak from Cargo Hatch on Barge G7- Mass Leak 3.12 g/s 32

Figure 3-14. Leak from Pressure Relief Valve on Barge G8- Mass Leak 5.78 g/s 33

Figure 3-15. Example #1 from LSI Ground Survey – Leaking Hatch on Barge L1- Total Mass Emission from Barge 0.521 g/s 36

Figure 3-16. Example #2 from LSI Ground Survey – Leaking Valve on Barge L2- Total Mass Emission from Barge 0.521 g/s 37

Figure 3-17. Example #3 from LSI Ground Survey – Leaking Hatch on Barge L6- Total Mass Emission from Barge 0.415 g/s 37

Figure 3-18. Example #1 from Ground Survey with LDEQ Camera – Leaking Hatch from Barge L10 (there was no OTM 10 emission flux measurement for this time period) 39

Figure 3-19. Example #2 from Ground Survey with LDEQ Camera – Leaking Vent from Barge L23- Total OTM 10 Mass Emission from Barge 0.490 g/s 40

iv


September 2009

List of Acronyms

AM Alkane Mixture

BEM Barge Emission Measurement

DQI Data Quality Indicators

ECPB Emissions Characterization and Prevention Branch

EPA U.S. Environmental Protection Agency

FAA Federal Aviation Administration

HLDS Hawk Leak Detection System

LDEQ Louisiana Department of Environmental Quality

LSI Leak Surveys Inc.

MDL Minimum Detection Limit

MOP Miscellaneous Operating Procedures

MSCHD Memphis and Shelby County Tennessee Health Department

NEdT Noise Equivalent Delta Temperature

NERL National Exposure Research Laboratory

NRMRL National Risk Management Research Laboratory

OAQPS Office of Air Quality Planning and Standards

OP-FTIR Open-Path Fourier Transform Infrared

ORD Office of Research and Development

ORS Optical Remote Sensing

OTM 10 EPA ORS Test Method OTM 10

PAC Path Averaged Concentration

PAMS Photochemical Assessment Monitoring Station

PGIE Passive Gas Imaging Equipment

Figure 3-20. Example #3 from Ground Survey with LDEQ Camera – Leaking Valve from Barge L13- Total OTM 10 Mass Emission from Barge 0.106 g/s 40

Figure 4-1. Screenshot from FLIR Camera Showing Leak from 9/24/ 2008, Event #1 44






v

PIC Path Integrated Concentration

QA Quality Assurance

QAPP Quality Assurance Project Plan

QC Quality Control

R4 EPA Region 4

R6 EPA Region 6

RSD Relative Standard Deviation

SOP Standard Operating Procedures

SSE Sum of Squared Errors

TCEQ Texas Commission on Environmental Quality

AM Total Hydrocarbon

TNMHC Total Non-methane Hydrocarbons

UV-DOAS Ultraviolet Differential Optical Absorption Spectroscopy

VOC Volatile Organic Compound

VRPM Vertical Radial Plume Mapping

WAM Work Assignment Manager

WC Wind Criteria

WSC Wind Speed Criteria


September 2009

vi


September 2009

Executive Summary

Recent airborne remote sensing survey data acquired with passive gas imaging equipment (PGIE), also known as infrared cameras, have shown potentially significant fugitive volatile organic carbon (VOC) emissions from petrochemical transport barges. A collaborative group with members from the United States Environmental Protection Agency Region 6 (EPA R6), the EPA Office of Research and Development (ORD) National Risk Management Research Laboratory (NRMRL) and National Exposure Research Laboratory (NERL), the Louisiana Department of Environmental Quality (LDEQ) and the Texas Commission on Environmental Quality (TCEQ) was formed to further investigate this topic. The common goals of the collaboration centered on improving knowledge of fugitive emissions from this source category and advancing field application information for select remote sensing techniques useful for identification and assessment of fugitive emissions from difficult to monitor sources such as barges.

To meet these goals the group conducted a field campaign in Baton Rouge, Louisiana from September 24 through October 9, 2008. This field campaign is described in this report and involved several complementary remote sensing and onboard leak rate measurement efforts. The study included aerial PGIE surveys of barges located on the Mississippi River and inter-coastal Waterway to identify barges with significant fugitive emissions. Additional ground-based PGIE observations of barges from the Port Allen Lock wall and also onboard a number of barges were conducted to closely observe fugitive leaks and identify leaking components. To support this work, an LDEQ study quantified emission leak rates using a bagging technique from a total of eight barges that were identified by the aerial remote sensing PGIE survey. To complement these efforts, EPA method OTM 10 with open-path Fourier transform infrared spectroscopy was used at the Port Allen lock to produce hydrocarbon emission measurements from barge traffic traveling through the lock.

The aerial PGIE survey detected leaks from 45 different barges located in the Mississippi River and the Intracoastal Waterway over a five day period. The ground-based PGIE monitoring detected leaks from over 18 different barges in the Port Allen lock during the study. The remote sensing surveys provided significant information regarding the practical use of infrared cameras for detection of emissions from petrochemical transport barges. This study produced a PGIE image database that informs the use of this technology by providing a basis for comparison of the qualitative PGIE leak images with estimated leak rates. This comparison helps improve PGIE survey technique understanding for barge emissions and other source categories.

vii


September 2009

In general, the employed PGIE equipment was found to be robust, easy to use, and possessed sufficient detection sensitivity for this application. The PGIE remote sensing approach was judged to be extremely useful for both aerial survey and close range fugitive leak inspection of petrochemical transport barges. The PGIE technique was able to identify a large range of leaks with large leaks detectable from the air and smaller leaks more easily observed at close range. PGIE observations were easier to execute during mid-day to late afternoon time periods due to more favorable background imaging conditions (improved background radiance from hot barge surfaces and lower shadow interference) and because fugitive emissions were likely more pronounced as the barges became heated by solar radiation and ambient temperature during the day. PGIE observations were very useful for identification of specific leaking components and verification of subsequent leak repair activities.

Based on aerial observations, eight barges with observed large leaks were selected for onboard leak emission rate measurements as part of the LDEQ bagging survey. For this effort, a total of 23 leak points from eight barges were bagged to estimate mass emission rates. The measured total non-methane hydrocarbon emissions flux values from individual leaks during the bagging study ranged from 0.07 g/s to 5.77 g/s. Summing all measured leaks for each individual barge yielded a barge total leak rate ranging from 1.13 g/s to 6.24 g/s. The average value of total leak rate measurement for eight barges was 3.3 g/s.

EPA method OTM 10 monitoring was conducted at the Port Allen lock wall from September 24 through October 9. A total of 97 barge sets passed through the lock during the observation period. Six barge events showed significant fugitive hydrocarbon emissions as measured by OTM 10 with values ranging from 0.047 g/s to 3.39 g/s alkane mixture (AM) flux rate with an average value of 0.83 g/s. The instrumentation used to apply the OTM 10 method exhibited sufficient operational robustness and detection sensitivity during the current study. Additionally, the OTM 10 technique was able to identify and assess emission rates from a range of leak sizes as long as the prevailing wind brought the emitted plume through the vertical plane of the measurement configuration.

Due to project constraints, there was no opportunity to conduct simultaneous emission measurements by OTM 10 and the bagging technique on the same barge. A baseline comparison of measurement results on different barges shows that the average total barge emission estimate by OTM 10 (0.83 g/s) was lower than the similar average from the bagging study (3.3 g/s). The maximum total barge emission estimates from the two techniques were more comparable (3.39 g/s with OTM 10 and 6.24 g/s with the bagging method). The somewhat lower OTM 10 values may be partially explained by the fact that the barges selected for the bagging experiments were identified by airborne survey as

viii


September 2009

having very significant leaks and may not represent an average emission case whereas the OTM 10 measurements were conducted on barges moving through the lock with no selection process and therefore may represent a more typical sample cross section.

Note that the emission estimates presented in this report represent a snapshot in time. Fugitive emissions from petrochemical barges are believed to vary significantly due to ambient temperature, thermal load, product mix, load state, and equipment condition and equipment design. Since there is limited information on how these variables affect fugitive emissions, extrapolation of data contained in this report is not recommended.

ix


September 2009

1. Introduction

1.1 Background

Recent airborne survey data using passive gas imaging equipment (PGIE), also known as infrared cameras, have shown potentially significant fugitive volatile organic compound (VOC) emissions from petrochemical transport barges. This is of interest as VOCs are precursors to ground-level ozone formation, contributing to the degradation of air quality, especially in urban areas. A collaboration of interested parties was formed to further investigate this issue. This group has common interests to expand knowledge of this source category and to further develop Optical Remote Sensing (ORS) techniques which facilitate fugitive emission identification and measurements from these and related sources.

The collaborative group consists of two main sub-groups which will individually sponsor and execute two separate Barge Emission Measurement (BEM) studies. Subgroup 1 consists of the United States Environmental Protection Agency Region 6 (EPA R6), the EPA Office of Research and Development (ORD) National Risk Management Research Laboratory (NRMRL) and National Exposure Research Laboratory (NERL), the Louisiana Department of Environmental Quality (LDEQ) and the Texas Commission on Environmental Quality (TCEQ). Subgroup 1 financially sponsored and executed BEM1 which occurred in Baton Rouge, Louisiana in September 24 through October 9, 2008 and is the subject of this report.

Subgroup 2 consists of EPA Region 4 (EPA R4) and the Memphis and Shelby County Tennessee Health Department (MSCHD) which was awarded an EPA Communities-Scale Monitoring Grant to plan and execute BEM2 in the fall of 2009. Each BEM project will benefit through active involvement of the above mentioned subgroups in addition to consultation from the EPA Office of Air Quality Planning and Standards (OAQPS), the BLF Consulting Group, and interested industry groups. The results of the studies will likely be compared in a separate publication.

1.2 Project Description

This report describes the BEM 1 field campaign conducted in Baton Rouge, Louisiana from September 24 to October 9, 2008. BEM 1 investigated VOC emissions from petro-chemical transport barges using portable gas imaging equipment PGIE (infrared cameras), EPA Method OTM 10 with Open-path Fourier transform infrared (OP-FTIR) spectrometers, in addition to leak bagging tests (manual leak rate measurements).

1


September 2009

The objectives of the study were:

• To improve knowledge of fugitive VOC emissions from petrochemical transport barges.

• To demonstrate and advance the field application of select ORS techniques (EPA OTM 10 OP-FTIR and PGIE) for identification and quantification of fugitive emissions from difficult to monitor sources.

• Identify sources of fugitive leaks from multiple barges

To accomplish these goals, the project team conducted several complementary efforts:

1. Aerial PGIE surveys of barges located on the Mississippi River and inter-coastal water ways identified barges with significant fugitive emissions.

2. Ground-based PGIE observations of barges from the Port Allen Lock wall and also onboard several barges identified and closely observe fugitive leaks.

3. Onboard leak emission bagging measurements were conducted by LDEQ on several barges to quantify leak rates and allow comparison with PGIE images.

4. EPA Method OTM 10 with open-path Fourier transform infrared spectroscopy was used at the Port Allen lock to produce hydrocarbon emission measurements from barge traffic traveling through the lock.

The body of this report summarizes the main aspects of the BEM 1 study with collections of representative images and emission measurement details contained in the Appendices A through I. With the exceptions noted below, this project was conducted by ARCADIS U.S., Inc. (ARCADIS), Durham, NC, under EPA ORD contract No. EP-C-04-023, Work Assignment No. 4-47. ARCADIS executed the OTM 10 portion of the field campaign, analyzed the OP-FTIR and OTM 10 data, produced draft versions of data tables and image collections and contributed to descriptions continued in this report. EPA Personnel were primary authors on the main body and summary sections of the report.

Section 2 of this report describes the measurement methods, instruments, and field setup for the BEM1 campaign.

Section 3 of the report describes the aerial and ground-based PGIE observations of barges on the Mississippi River and Intracoastal Waterway and Port Allen Lock. This portion of the

2


September 2009

study was funded by EPA and was executed by Leak Surveys, Inc. (LSI), under subcontract to ARCADIS and by ARCADIS using PGIE owned by LDEQ. A summary of LSI results along with several representative PGIE screenshots are presented in Sections 3.1 (aerial) and 3.2 (ground-based). Section 3.3 presents this same information for the PGIE observations made on the lock wall by ARCADIS. Additional details and images from this part of the project are contained in Appendices A through E.

Section 4 of the report describes measurements made on the Port Allen Lock wall by ARCADIS using two scanning optical remote sensing instruments (OP-FTIR), in combination with the OTM 10 protocol (see http://www.epa.gov/ttn/emc/prelim.html). These measurements provided hydrocarbon emission estimates of a representative alkane mixture (AM) and speciated concentration measurements for several trace compounds for barges passing through the lock. Section 4 presents a description of notable events, the AM mass emission flux values measured during each event, a screenshot of the leaks from the PGIE observations, and the results of the trace compound analysis. Additional information on the OTM 10 portion of the study is contained in Appendices F and G.

Section 5 summarizes the findings of the onboard leak rate measurements performed by Sage Environmental for LDEQ on several barges during the study. The leak bagging measurements used U.S. EPA Protocol for Equipment Leak Emission Estimates (U.S. EPA, 1995), with some variations. The LDEQ report on the bagging experiments is reproduced in its entirety in Appendix H for reference. It is noted that a draft version of this BEM 1 study report was reviewed by the American Waterways Operators and their comments concerning the LDEQ bagging study along with responses from Sage Environmental Consulting are reproduced in Appendix J.

Section 6 of the report presents a general comparison of emissions levels from the set of barges passing through the Port Allen lock observed during the OTM 10 study with the barges measured during the LDEQ bagging study. Section 7 provides QA information including discussions on general uncertainty and data limitations. Section 8 summarizes the conclusions for the study.

This report has been reviewed by the Office of Research and Development, U.S. EPA, and approved for publication. Approval does not signify that the contents necessarily reflect the views and policies of the agency nor does mention of trade names or commercial products constitute endorsement or recommendation for use.

3

http://www.epa.gov/ttn/emc/prelim.html


September 2009

2. Description of Test Sites, Measurement Methods and Site Deployment

The experimental approach for BEM1 included three main elements: PGIE for aerial and ground-based observations, fugitive emission estimation by EPA Method OTM 10 at the Port Allen and leak bagging tests of emissions from several barges. Following a general description of PGIE, Sections 2.1 and 2.2 describe the details of the aerial and ground-based PGIE observations conducted for this study. Section 2.3 shows the U.S. Army Corps of Engineers Port Allen lock site and describes the EPA OTM 10 method used to assess the mass emission flux of a representative alkane mixture (AM) in addition to trace compound speciation. Section 2.4 describes the bagging procedure employed in the LDEQ effort to quantify fugitive emissions leaks onboard several barges.

Of special interest to this study is the use of PGIE remote sensing systems to investigate fugitive emissions from barges. The PGIE infrared cameras were used to qualitatively detect the presence of fugitive emissions and to observe the leaking component to inform emissions inventory knowledge. The details of the specific PGIE used in this study are provided in Section 2.1. In general, the infrared camera detects thermal energy emitted by objects in the optic field of view or scene as motion imagery or video. Thermal energy is absorbed by molecules in the camera’s field of view. If the molecules are present in high concentrations, and if their infrared-active molecular vibrations are within the bandpass of the camera, the molecules are detected. The fugitive emission or leak is detected by the PGIE operator by observing the relative brightness of the camera pixels that comprise the scene. Gas leaks appear as black or very dark plumes in the video relative to other objects in the scene. These plumes are dynamic as well, which assists in discriminating gas leaks from other scene objects such as thermal shadows or cold objects. The camera video is recorded onto a solid state recorder for future analysis. More information on the PGIE camera can be found in the Texas Commission on Environmental Quality SOP #SAMP-020, Operation of FLIR Systems THERMAGAS GasFindIR Camera, presented as Appendix C of the project Quality Assurance Project Plan (QAPP) (EPA, 2008).

The study was conducted from September 24 through October 9, although measurements from each of the three study elements were not collected continuously during this time. The weather conditions observed during the study period were relatively normal conditions for the Port Allen area (normal average daytime high temperatures ranging from 86° F on September 24 to 82°F on October 9).

4


September 2009

2.1 Aerial PGIE Observations

The aerial PGIE observations were made by Leak Surveys Inc. (LSI) from September 24 through October 1, 2008 (10 am to 5 pm). The crew used the Hawk Leak Detection System (HLDS), developed by LSI after 12 years of research and development on optical imaging.

The PGIE used was a modified Indigo (FLIR/Indigo Systems Corp., Goleta, CA) Merlin MID camera, which is a specialized thermal imaging camera that allows the operator to visualize a plume of VOC gases, allowing the leaking barge component to be identified. The PGIE was mounted on a helicopter using a FAA inspected and certified Tyler vibration isolating mount. The camera video was cabled to the operator inside the aircraft. The aircraft was generally deployed to conduct monitoring of barges upriver of the lock prior to actual arrival at the lock site. A standard digital camera was used for photographing the barge under surveillance for future reference. A GPS unit was used to log the location and time and date of the contact.

The PGIE has a nominal spectral range of 1 to 5.4 µm. Using a 30 × 30 µm InSb detector with a 320 × 240 pixel array, the camera has the capability to vary the integration times from 5 ms to 16.5 ms. The detector is operated at near liquid nitrogen temperatures using an integral Sterling cooler which provides the system with an NEdT of no more than 18 mK providing excellent sensitivity.

The spectral range is further limited with the use of a notch filter specifically designed for the detection of hydrocarbon infrared adsorption in the 3 micron region. The narrow bandpass range of the filter is less than the infrared spectral absorption of gas-phase hexane. The filter notch is positioned so that alkane gases have a significant response within the bandpass range.

Various lenses including a 25 mm, a 50 mm, and a 100 mm lens were used. The 25 mm lens provided a 22 × 16 degrees field of view with an f-number of 2.3. The 50 mm lens provided an 11 × 8 degrees field of view with an f-number of 2.3.

The use of a narrow bandpass filter provides spectral discrimination that allows the detection of compounds that have a vibration mode in the infrared region of the filter. Not all hydrocarbons have infrared absorptions within the filter range. Table 2-1 shows the theoretical relative response of various compounds of interest using 1 cm-1 resolution infrared spectra (Infrared Analysis, Inc., Anaheim, CA). Using propane as the reference spectrum with a relative response of 100, methane’s response is approximately 10 percent of the same concentration of propane and hexane is 1.5 times the response of propane at

5

Table 2-1. Relative Response of Hydrocarbon with LSI Infrared Imaging Camera (FLIR)

Relative Response Compound Propane = 100%

Methane 9

Ethane 43

Propane 100

Butane 118

Iso-Butane 137

Pentane 143

Hexane 155

Heptane 157

Octane 136

Ethylene 3

Propylene 20

Iso-Butylene 37

2-Methyl-2-Butane 4

1-Pentene 7

2-Methyl-2-Pentene 7

Benzene 4

Toluene 21

o-Xylene 38

p-Xylene 23

m-Xylene 32


September 2009

the same concentration. The filter is set to the infrared region of the spectrum that primarily corresponds to the infrared absorption of alkanes. Other hydrocarbons exhibit various degrees of absorption of infrared energy in this region as indicated in the table.

The aerial surveys were performed using a two-person crew consisting of the pilot and the camera operator. The survey was conducted by focusing the PGIE on the river and searching for barge leaks. If a leak was found, the pilot circled back above the source and the camera operator recorded the leaking emissions for a period of approximately 2 minutes. The results of the aerial survey are presented in Report: Leak Detection using LSI Infrared Gas Imaging, LDEQ Barge Study (27 October 2008) which is included as Appendix A.

6


September 2009

2.2 Ground-based PGIE Observations

The ground-based PGIE observations at the lock were made by LSI (September 24 and September 28-30, 2008) using a modified Indigo (FLIR/Indigo Systems Corp., Goleta, CA) GasfindIR MID camera and by ARCADIS using the LDEQ FLIR camera (September 28 and October 1-9, 2008). There were no FLIR observations made at the lock on September 2527, 2008. LSI also made the observations at the onboard several barges during the LDEQ bagging study (September 24-28, 2008). The PGIE used to perform these optical imaging leak surveys was similar to the camera used for the aerial surveys described in Section 2.1.

The light weight and small size of the PGIE allowed it to be hand-carried for ground observations of barges. Leaking components on the barge (if present) were identified and logged. The potential source of the leak was identified (i.e., hatch cover, pressure relief valve) and the position on the barge (i.e., cargo tank #4) was determined.

The results of the LSI ground survey are presented in Report: Leak Detection using LSI Infrared Gas Imaging, BEM1 Barge Study; Ground Crew Survey (21 October 2008) which is included as Appendix C.

Images from the LSI and ARCADIS ground PGIE observations are presented in Appendices D and E.

2.3 Scanning OP-FTIRs and OTM 10 Protocol

Two scanning open path Fourier transform infrared (OP-FTIR) spectrometers, in combination with the OTM 10 protocol (see http://www.epa.gov/ttn/emc/prelim.html), were used to provide alkane mixture (AM) emission flux and speciated measurements for nine trace compounds (methane, methanol, benzene, ethylene, acetylene, propylene, propane, ethane, and carbon monoxide).

Supporting measurements for this phase of the study included meteorological data and distance measurements. Testing and measurement protocols included:

• PIC emission measurements with two OP-FTIR instruments

• Meteorological data collection with the R.M. Young heads

• Optical path length determination with a Topcon theodolite

7

http://www.epa.gov/ttn/emc/prelim.html


September 2009

• Calculation of AM emission flux using Flux Calc (ARCADIS software employing the VRPM method).

Figure 2-1 shows an overhead view of the U.S. Army Corps of Engineers Port Allen lock study site, with the approximate locations of the project measurement configurations. Two 3beam OTM -10 configuration planes were deployed along the southern edge of the lock (denoted by the red lines, not to scale vertically) using one EPA and one ARCADIS scanning OP-FTIR. The end of the two planes was defined by one common scissor lift (tower in the middle), which was used to mount the two elevated mirrors of each OTM 10 configuration. The lowest mirror in each configuration was deployed on the surface of the lock wall walkway (0.1 m height) and the elevated mirrors were positioned at heights of approximately 3 m and 6 m above the walkway. The locations of the two scanning OP-FTIR instruments were near each end of the lock as indicated in the figure. The length of the EPA and ARCADIS OP-FTIR plane configurations were 169 m and 153 m, respectively. The overall length of the lock from gate to gate was approximately 360 m and the width of the lock was 25 m. The OTM 10 planes were located approximately 1 m away from the inside lock wall edge. Additional images illustrating the OTM 10 deployment and barge traffic in the lock are contained in Figures 2-2 and 2.4.

Two additional optical beam paths were deployed (one from each OP-FTIR instrument) across the surface of the lock to collect supplemental data on alkane mixture and trace VOC concentrations. Although the project Quality Assurance Project Plan stated that these data would be collected with ultraviolet differential optical absorption spectroscopy (UV-DOAS), the project team determined that the OP-FTIR data could be analyzed for trace VOC concentrations, so the UV-DOAS instrument was not deployed at the site due to limited project resources and eye safety concerns for lock personnel.

Originally, it was anticipated that the OTM 10 configuration would be deployed along the northern edge of the lock. However, at the time of the field campaign, the winds were largely from the north, and the configuration was deployed on the southern edge of the lock. The prevailing winds at the site during the measurements are denoted by the wind rose in the lower left hand corner of Figure 2-1. Note that in order for the OTM 10 configuration to measure fugitive emission from a particular barge, the wind vector must have a significant component from the north in order for the emitted plume to traverse the OTM 10 measurement plane.

OP-FTIR data were collected with each configuration from September 24 through October 9, 2008. Data from the barge traffic were recorded including the time of entry and exit from lock, reported cargo from the U.S. Army Corps of Engineers traffic log, and visual inspection of

8


September 2009

cargo labels on each barge. Emissions flux values were calculated for each event by summing the flux values measured from each configuration.

Figure 2-1. U.S. Army Corps of Engineers Lock Study Site and Measurement Configurations

9


September 2009

Figure 2-2. Images Illustrating OTM 10 Setup and Barges at the Port Allen Lock

2.3.1 OTM 10 and the Vertical Radial Plume Mapping Method

The following is a general description of the ground-based barge measurements at the U.S. Army Corps of Engineers Port Allen lock. For this phase of the campaign, two OP-FTIR instruments were placed around the lock area to execute a modified version (3-beam) of EPA Method OTM 10 to quantify the mass emission flux of and alkane mixture (AM) from the barges located in the lock.

The project used two 3-beam OTM 10 flux measurement configurations to quantify the fugitive emissions from the barge. The Vertical Radial Plume Mapping (VRPM) method is the analytical part of the OTM 10 flux measurement and generally discussed in EPA OTM 10 Optical remote sensing for emission characterization from non-point sources, which describes direct measurement of pollutant mass emission flux from area sources using ground-based optical remote sensing (ORS). The OTM 10 technique utilizes open-path spectroscopic instrumentation to obtain path-integrated pollutant concentration information along multiple optical paths. The multi-path pollutant concentration data along with wind vector information are processed with a plane-integrating VRPM computer algorithm to yield

10


September 2009

a mass emission flux for the source. Figure 2-3 shows a general 5-beam TOM 10 VRPM measurement configuration. For this project, a 3-beam configuration was used having no intermediate mirrors with 3 beams extending along the ground (mirror deployed on surface of top of the lock wall), middle, and top scissor lift positions.

Figure 2-3. General OTM 10 VRPM Measurement Configuration

The VRPM computer algorithm uses a smooth basis function minimization routine of a bivariate Gaussian function to generate mass emission flux information from species concentration and wind data. For this measurement campaign, the VRPM configuration utilized a three-beam configuration which leads to a reduced form of the bivariate Gaussian in polar coordinates (r, θ). The standard deviation in the crosswind direction is assumed to be about four times the length of vertical plane (r1).

11

A ⎪⎧ 1 ⎡(r ⋅ cosθ − 12 r1 )2 (r ⋅ sinθ − mz )2 ⎤⎪⎫

G(A,σ z , mz ) = exp⎨− ⎢ + ⎥⎬ (1)2π (4r )σ ⎪ 2 (4r )2 σ 2 ⎪1 z ⎩ ⎣ 1 z ⎦⎭

i⎛ r ⎞

2

SSE(A,σ z , mz ) = ∑⎜⎜PACi − ∫G(ri ,θ i , A,σ mz

)dr / ri ⎟⎟ (2)

i ⎝ 0 ⎠


September 2009

Where:

A = normalizing coefficient, adjusts for the peak value of the bivariate surface;

mz = peak location in Cartesian coordinates;

σz = vertical standard deviation in Cartesian coordinates;

r1 = length of VRPM plane;

A, mz, and σz are the unknown parameters to be retrieved by the fitting procedure. An error function (sum of squared errors, SSE) for minimization is defined as:

Where PACi is the measured path-averaged concentration (PAC) value for the ith beam. The SSE function is minimized using the Simplex method to solve for the three unknown parameters. This process is for determining the vertical gradient in concentration. It allows an accurate integration of concentrations across the vertical plane as the long-beam ground level PAC provides a direct integration of concentration at the lowest level.

Once the parameters of the function are found for a specific run, the VRPM procedure calculates the concentration values for every square elementary unit in a vertical plane. Then, the VRPM procedure integrates the values, incorporating wind speed data at each height level to compute the flux. This enables the direct calculation of the flux in grams per second (g/s), using wind speed data in meters per second (m/s). Further information on the VRPM method for fugitive source emission measurements in general can be found in Thoma 2005, U.S. EPA 2007a with specific details of this deployment in U.S. EPA 2007b.

This measurement project has several unique features regarding the use of EPA Method OTM 10 which is typically used for ground-level area source measurements using a 5-beam approach. Specifically, this field study utilized a 3-beam approach and the source was not the typical ground level area source. The 3-beam OTM 10 approach was chosen for this project since it was much more important to obtain a larger number of measurement cycles while the mobile source was contained in the lock rather than a fewer number of cycles with a five beam approach since the horizontal spatial location of the plume was not of primary importance. In analyzing the PIC data using the 3-beam approach, several assumptions are required. The peak plume concentration was assumed to be centered along the crosswind axis of the OTM 10 configuration, and the σy parameter (horizontal dispersion coefficient) of

12


September 2009

the measured plume was assumed to be equal to ½ the length of the OTM 10 configurations. It was necessary to make these assumptions because the 3-beam OTM 10 approach does not include two intermediate surface beam paths which are used to obtain information on the horizontal location and dispersion of the plume. Section 7.3 has contains a discussion of uncertainty associated with these assumptions.

The lock wall configuration was not a typical area source deployment however it was assumed that the emission form the barges acted similarly to a close-coupled area source. The emitted plumes from the barges were assumed to be initially small in spatial extent but would experience significant dispersion by eddy mixing before exiting the lock and passing through the OTM 10 plane. This is likely since the barges were significantly below the lock wall top (approximately 7m to 12 m) and the emitted plumes could experience several dispersive/mixing mechanisms (such as stagnation, turbulence, channeling) depending on ambient wind direction and speed so the plumes could evolve more than in a flat wind swept scenario with similar downwind standoff. This results in a relatively well-developed plume exiting the lock and being transported by the free-flowing winds to the OTM 10 plane. This is illustrated in Figure 2.4 and these assumptions are further discussed in Section 7.3. The distance range below the lock wall (≈ 7 m to 12 m) reflects the approximate lock operation water level height change during the study.

≈ 7 to 12 m

6 m

≈ 25m

≈ 7 to 12 m

≈ 25m

OOTTMM 1100 FluFluxx PPllaneane

Lock Wall

Barge

Dispersion ofPlume byEddy Mixing

Free-Flowing Winds

Emission plume

Lock Wall

Barge

≈ 7 to 12 m

6 m

≈ 25m

Dispersion of Plume by Eddy Mixing

Free-Flowing Winds

Emission plume

Figure 2-4. Representation of Lock Cross Section and Wind Flow

13


September 2009

2.3.2 OTM 10 Fugitive Emission Quantification

The scanning OP-FTIR measurement system and associated OTM 10 planes employed a default configuration as described above. The default dwell time for each mirror was 30 seconds. In general, the OP-FTIR spectrometer is designed for both fence-line monitoring applications and real-time, on-site, remediation monitoring and source characterization. The OP-FTIR instrument consists of an infrared light beam, modulated by a Michelson interferometer. The infrared beam is transmitted from a single telescope to a retro-reflecting mirror target, which is usually set up at a range of 100 to 500 m. The returned light signal is received by the single telescope and directed to a detector. The light is absorbed by the molecules in the beam path as the light propagates to the retro-reflecting mirror and again as the light is reflected back to the analyzer. The advantage of OP-FTIR monitoring is that the concentrations of a multitude of infrared absorbing gaseous chemicals can be detected and measured simultaneously, with high temporal resolution. Figure 2-5 presents a picture of the OP-FTIR instrument.

2.3.3 Supporting Measurements for Ground-based ORS

2.3.3.1 Meteorological Data

Meteorological data including wind direction and wind speed were continuously collected during the measurement campaign with two R.M. Young model 05103 meteorological heads. The instrument is automated and collects real-time data from its sensors and records time-stamped data, which are transmitted to a desktop computer via a radio frequency modem R. M. Young model 32500. The meteorological heads were deployed to collect wind speed and wind direction data during the study. As part of each VRPM configuration, one head was deployed on the surface of the lock wall at a height of approximately 3 meters, and the other head was deployed on top of the scissor lift platform at a height of approximately 6 meters above the lock wall.

More information on deploying R.M. Young meteorological heads can be found in MOP 6803 “Guidance for Deploying and Using ORS Supplemental Instrumentation” of the Emissions Characterization and Prevention Branch (ECPB) ECPD Optical Remote Sensing Facility Manual (U.S. EPA, 2004).

14


September 2009

Figure 2-5. IMACC OP-FTIR Instrument and Scanner

2.3.3.2 OP-FTIR Instrument-Mirror Distance

The physical distance between the ORS instruments and the mirrors was measured using a Topcon, Inc. model GTS-211D theodolite. More information on setting up and operating the theodolite can be found in MOP 6822 “Determining the Geographical Locations of the ORS Measurement Locations” of the ECPD Optical Remote Sensing Facility Manual (U.S. EPA, 2004).

2.3.4 PIC Emission Measurements with OP-FTIR Instrument

To calculate the mass emission flux using the OTM 10 method, the acquired OP-FTIR data must be analyzed to produce a PIC value. For this project, OP-FTIR data reduction focused on the PIC values of a representative alkane mixture (AM) by spectroscopic analysis of the infrared absorption features in the C-H stretch spectral region around 2950 cm-1. The analysis focused on an alkane mixture (butane, pentane, hexane, heptane, octane, nonane, decane) since performing spectral analysis of each individual compound is not possible due to the similarity in the shapes of the absorption bands. Additionally, the molecular weight of the target compound is necessary to calculate the mass emission flux using the OTM 10 method. Spectroscopic analysis of the AM also yielded the average molecular weight of AM for each concentration determination. More information on the method used for spectroscopic analysis of the AM can be found in Appendix F of this report.

15


September 2009

In addition to the AM quantification, individually quantifiable hydrocarbons species (e.g., methane) were analyzed if present at concentrations above the MDL for the OP-FTIR.

The general measurement of the Path Integrated Concentration (PIC) of individually identifiable analyte gases using the IMACC OP-FTIR instrument is described in MOP 6808 “Multiple-Path Data Collection Using a Scanning IMACC Monostatic OP-FTIR” of the ECPD Optical Remote Sensing Facility Manual (U.S. EPA, 2004). A detailed description of the procedure used for PIC concentration analysis with IMACCQuant software is contained in MOP 6827 “Procedures for OP-FTIR Concentration Data Analysis Using IMACCQuant Software”. The estimated minimum detection levels for the target analytes of the OP-FTIR instrument are presented in Table 2-2.

Table 2-2. Target Compound List

OP-FTIR Estimated Detection Limit Compound for Optical Path Length = 300 m,

1 min. averaging (ppb)

Alkane Mixture (AM) 2

Methane 2

Methanol 4

Benzene 20

Ethylene 1

Acetylene 2

Propylene 4

Propane 10

Ethane 10

2.3.5 Meteorological Data Collection with the R.M. Young Heads

Meteorological data including wind direction and wind speed were continuously collected during the sampling/measurement campaign with an R.M. Young Model 05103 meteorological head. The instrument is automated and collects real-time data from its sensors and records time-stamped data, which are transmitted to a desktop computer via a transmitter.

For this project, a wind direction and speed-sensing head was used to collect data at heights of approximately two and six meters above ground. The sensing head for wind direction incorporates an auto-northing function (automatically adjusts to magnetic north) that

16


September 2009

eliminates the errors associated with subjective field alignment to a compass heading. The sensing heads incorporate standard cup-type wind speed sensors. Post-collection, a linear interpolation between the two sets of data is done to estimate wind velocity as a function of height. More information on deploying and operating the R.M. Young Model 05103 meteorological instrumentation can be found in MOP 6803 of the ECPD Optical Remote Sensing Facility Manual (U.S. EPA, 2004).

2.3.6 Optical Path Length Determination with the Topcon Theodolite

The physical distance between an OP-FTIR instrument and a mirror was determined by measurement with a Topcon, Inc. model GTS-211D theodolite. The instrument manufacturer certifies the instrument accuracy and precision to better than 1 cm. Azimuth and elevation angles can also be determined using the theodolite. The measurement is a manual operation, with the results recorded by hand, and is followed by transcription to a spreadsheet for data archiving and calculations.

Due to folding of the optical beam by the mirror, the optical path length is twice the physical distance between the instrument and mirror. In other words, the optical beam passes through the physical path twice. More information on deploying and operating the Topcon theodolite can be found in MOP 6822 of the ECPD Optical Remote Sensing Facility Manual (U.S. EPA, 2004).

2.3.7 Calculating Emission Flux using the VRPM Method

The calculated emission flux is generated by inputting the measured PIC data into the VRPM algorithm. The algorithm is performed using Matlab software (MathWorks, Inc., Natick, MA). The VRPM method maps the concentrations in the plane of the measurement. The horizontal dimension of this plane is defined as the distance between the OP-FTIR instrument and the most distant mirror used in the configuration. The vertical dimension of this plane is defined as the distance from the surface to the point where the extrapolated concentration values (extrapolated based on the vertical concentration gradient) approaches zero. This height is not determined until the data are processed in the VRPM algorithm. By scanning in a vertical plane downwind from an area source, one can obtain plume concentration profiles and calculate the plane-integrated concentrations. The flux is calculated by multiplying the plane-integrated concentration by the wind speed component perpendicular to the vertical plane. The flux leads directly to a determination of the emission rate (Hashmonay et al., 1998; Hashmonay and Yost, 1999; Hashmonay et al., 2001; Thoma et al. 2005). More information on the procedures used to generate the plume maps and the calculated emission flux can be found in MOP 6842 “Using the vertical Radial Plume

17


September 2009

Mapping (VRPM) Configuration with Wind Data to Create Plume Concentration Profiles and Calculate Emission Fluxes” of the ECPD Optical Remote Sensing Facility Manual (U.S. EPA, 2004).

2.3.8 Site Deployment Description at the Lock

The prevailing winds during the time of the field campaign were largely from the north, and the dual OP-FTIR configuration was deployed on the southern edge of the lock. Figure 2-1 shows the VRPM configurations at the U.S. Army Corps of Engineers lock. Following ORS instrument setup and Data Quality Indicator (DQI) checks, ORS data acquisition began, continuously, for a period of several hours each day. During the surveys, measurements were taken along each optical path length (mirror) sequentially. The averaging time for each optical path was thirty seconds. Emissions flux values were calculated for each event by summing the average flux values measured from each configuration.

Data from barge traffic were recorded including time of entry and exit from lock, reported cargo from the U.S. Army Corps of Engineers traffic log, and visual inspection of cargo labels on each barge. The ground-based PGIE (FLIR camera) measurements were conducted simultaneously with the ORS measurements for a period of several hours each day, for all but three days of the sampling campaign. It is noted that Barge traffic through the Port Allen lock was lighter than usual during the study due to repair activities on the Intracoastal Waterway Bayou Sorrel Bridge, which was damaged by a tug boat accident just prior to the start of the field campaign.

2.4 Bagging Tests

The onboard-barge leak bagging tests were performed by SAGE Environmental Consulting for LDEQ from September 24-28, 2008. Some of the results from this study are presented in Sections 5 and 6 for comparison purposes with the full LDEQ report reproduced as Appendix H for reference. For the bagging measurements, SAGE followed the vacuum method described in the U.S. EPA Protocol for Equipment Leak Emission Estimates (U.S. EPA, 1995), with some variations. For compositional analysis, samples were collected by LDEQ in aluminum Summa canisters. A maximum of one canister was filled for each point tested. One canister was sometimes used for multiple sampling points in the same product service on the same barge. The LDEQ laboratory did the analysis using EPA PAMS analysis by GC/FID. Figure 2.6 shows images from the LDEQ leak bagging study.

18


September 2009

Figure 2-6. Images from LDEQ Leak Bagging Study

19


September 2009

3. Aerial and Ground-based PGIE Results and Discussion

The aerial and ground-based PGIE observations of barges on the Mississippi River and Intracoastal Waterway are summarized in this section. Representative PGIE snapshots are presented in Sections 3.1 (aerial) and 3.2 (ground-based). Section 3.3 presents this same information for the PGIE observations made in the lock on several other days by ARCADIS, with the LDEQ camera. No PGIE observations were made at the lock on September 25-27, 2008. Note that it is not possible to reproduce the details of a visible leak (as captured with video footage) using a single representative snap shot as required for this report.

3.1 Aerial PGIE Observations by LSI

The aerial (helicopter) PGIE observations were made by Leak Surveys, Inc. (LSI) from September 24 through September 30, 2008. The helicopter was airborne for approximately 6 hours each day. The LSI aerial crew dataset contains movies showing leaks from a total of 45 different barges located in Mississippi River and Intracoastal Waterway. Table 3-1 lists the LSI snapshot identification number, barge number, and a description of the suspected source of the leak(s). Additional information is contained in Appendices A and B.

Table 3-1. Summary Table of Barge Leaks Identified by LSI Aerial Survey

Filename Date Part Leaking Barge #

L000 9/24/2008 Two Large Valve Settings Towards Aft Side A1

L001 9/24/2008 Vent Stack at Bow of Barge A2

L002 9/24/2008 Vent Stack at Bow of Barge A3

L003 9/24/2008 Top Loading Hatches A4

L004 9/25/2008 Top Loading Hatches at Placid Refinery A5

L005 9/25/2008 Top Loading Hatches to the Aft of Barge A6

L006 9/25/2008 Top Loading Hatches at Bow of Barge A7



L009 9/26/2008 Top Loading Hatch at Bow of Barge A10

L010 9/26/2008 Top Loading Hatches and Vent A11

L011 9/26/2008 Top Loading Hatches at Aft side of Barge at Placid Refinery A12

L012 9/26/2008 Top Hatches on Barge A13

L013 9/26/2008 Top Hatches on Barge A14

20


September 2009

Filename Date Part Leaking Barge #

9/26/2008 Top Hatches and Vent Stack on Barge A15

L014 9/26/2008 Top Hatches and Vent Stack on Barge A16

9/26/2008 Top Hatches on Barge A17

9/26/2008 Top Hatches on Barge A11

L015 9/27/2008 Vent Stack on Barge A18

L016 9/27/2008 Repeat of Video 003, Boarded by Bagging Team A4

L017 9/27/2008 Top Loading Hatches at Placid Refinery A19

L018a 9/27/2008 Top Hatches on Barge A20

L018b 9/27/2008 Top Hatches on Barge A21

L018c 9/27/2008 Top Hatches on Barge A22

L019 9/27/2008 Top Hatches at TT Barge Cleaning Facility --Refilm A13

9/27/2008 Top Hatches at TT Barge Cleaning Facility A14

L020 9/27/2008 Vent Stack on Barge --Refilm A23

9/27/2008 Vent Stack on Barge A24

L021 9/27/2008 Hatches at Aft Side --in Intracoastal Waterway A25

9/27/2008 Vent at Aft Side --in Intracoastal Waterway A26

L022a 9/27/2008 Vent at Aft Side A27

L022b 9/27/2008 Vent Stack on Bow of Barge A28

L023a 9/27/2008 Center Vent on Barge A29

L023b 9/27/2008 Cent Hatch on Barge A30

L024 9/28/2008 Bow Hatch and Deck Hatch on Barge A31

L025 9/28/2008 Two Aft Hatches and one Side Hatch A32

L026 9/29/2008 Top Hatches on Barge --in Intracoastal Waterway A33

L027 9/29/2008 Vent Stack in Center of Barge --in Intracoastal Waterway A34

L028 9/29/2008 Vent Stack on Bow of Barge --Refilm A1

L029 9/29/2008 Vent Stack on Bow of Barge A35

L030 9/29/2008 Vent Stack on Bow of Barge --Refilm A36

L031a 9/29/2008 Top Hatches on Barge --Across from Locks A37

L031b 9/29/2008 Top Hatches on Bow --Across from Locks A31

L032 9/30/2009 Vent at the Bow of Barge --North of Locks A23

L033 9/30/2009 Forward Bow Hatch on Barge --in Intracoastal Waterway A38

21


September 2009

Figures 3-1 through 3-3 show three example snapshots from the LSI Aerial survey illustrating the type of leaks that were detected: Example #1 shows a vent on the aft side of Barge A27; Example #2 shows a vent stack in the center of Barge A34 from the Intracoastal Waterway; and Example #3 shows the top hatches on the bow of Barge A31across from the lock. Appendix B contains a larger collection of images providing information on various types of leaks identified during the helicopter survey.

Note that it is impossible to represent visual acuity of observed leaks with single image snapshots reproduced in this report. The leaks as viewed in moving video images are much more pronounced and generally easier to identify, particularly for small leaks.

Figure 3-1. Example #1 from LSI Helicopter Survey – Leak from Vent on Aft Side of Barge A27

22


September 2009

Figure 3-2. Example #2 from LSI Helicopter Survey – Leak from Vent Stack in Center of Barge A34

Figure 3-3. Example #3 from LSI Helicopter Survey – Leak from Top Hatches on Bow of Barge A31

23


September 2009

The LSI aerial surveys provided the following general conclusions about this remote sensing approach for fugitive emission detection from petrochemical transport barges:

1. The deployed PGIE equipment and airborne platform exhibited sufficient mobility, operational robustness, and detection sensitivity and are judged to be extremely useful for airborne remote monitoring of this type. This conclusion agrees with previous similar studies using this technology.

2. The PGIE technique was able to easily identify apparent large leaks from the air with smaller leaks somewhat more difficult to identify likely requiring an expert operator.

3. Ground surveys indicated that the leaks seen from the air were many times composed of numerous individual leaks upon closer inspection.

4. In all studied cases, leaks identified from the air were verified as being VOC leaks of significant volume upon ground inspection.

5. Aerial observations were easier to execute during mid-day to late afternoon time periods due to more favorable background imaging conditions (improved background radiance from hot barge surfaces and lower shadow interference) and because fugitive emissions were likely more pronounced as the barges became heated by solar radiation and ambient temperature during the day.

3.2 On board Barge PGIE Observations by LSI

LSI conducted ground-based PGIE observations on board eight barges in conjunction with the LDEQ bagging studies (September 24-28, 2008). The barges where monitoring occurred represented a subset of those identified by the aerial survey (Section 3.1) having large apparent leaks as viewed from the air. The purpose of the onboard PGIE observations was to provide close-up views of leaks prior to and during the bagging measurements. Table 3-2 lists the LSI filename number, the date of the observation, a description of the leak source, and the barge number or other descriptor.

24


September 2009

Table 3-2. Summary Table of Barge Leaks Identified by LSI Ground-based Observations

Filename Date Part Leaking Barge # or Other Descriptor

7 9/24/2008 Hatch Barge G1








15 9/25/2008 Ullage Hatch Barge G2 #2 Port Lower

16 9/25/2008 Cargo Hatch Barge G2 #2 Port

17 9/25/2008 Butterworth Hatch Barge G2 #2 Starboard Middle

18 9/25/2008 Butterworth Hatch Barge G2 #2 Port Middle

19 9/25/2008 Butterworth Hatch Barge G2 #1 Starboard Lower

20 9/25/2008 Alarm Test Rod Barge G2 #2 Starboard

21 9/25/2008 Butterworth Hatch Barge G2 #1 Port Lower

22 9/25/2008 Alarm Test Rod Barge G2 #2 Port

23 9/25/2008 Cargo Hatch Barge G2 #1 Starboard

24 9/25/2008 Butterworth Hatch Barge G2 #1 Starboard Middle

25 9/25/2008 Butterworth Hatch Barge G2 #1 Port Middle

26 9/25/2008 Butterworth Hatch Barge G2 #1 Starboard Upper

27 9/25/2008 Butterworth Hatch Barge G2 #1 Port Upper

28 9/25/2008 Butterworth Hatch Barge G2 #2 Port Lower

29 9/25/2008 Overview of Leaks Barge G2

30 9/25/2008 Bagging Process Showing Gas Venting Through Dry Gas Meter

31 9/25/2008 Vent Barge G3

32 9/25/2008 Pressure Relief Valve Barge G3

33 9/25/2008 Butterworth Hatch Barge G3 #1 Port Forward

34 9/25/2008 Ullage Hatch Barge G3 #1

35 9/25/2008 Butterworth Hatch Barge G3 #1 Port Aft

36 9/25/2005 Cargo Hatch Control Valve Barge G3 #1

37 9/25/2008 Butterworth Hatch Barge G3 Starboard Forward

25


September 2009




40 9/25/2008 Butterworth Hatch Barge G3 #3 Starboard Forward


42 9/25/2008 Butterworth Hatch Barge G3 #3 Port Aft


44 9/26/2008 Ullage Hatch Barge G4 #1 Port & #1 Starboard

45 9/26/2008 Both Hatches & Valve Barge G4 #2 Port

46 9/26/2008 Both Hatches & Valve Barge G4 #2 Starboard

47 9/26/2008 Ullage & Cargo Hatches Barge G4 #3 Starboard

48 9/26/2008 Cargo Hatch Control Valve Barge G4 #3 Port


50 9/26/2008 Ullage Hatch Barge G5 #1 Starboard

51 9/26/2008 Ullage Hatch & Valve Barge G5 #1 Port

52 9/26/2008 Ullage Hatch Barge G5 #2 Port





57 9/26/2008 Same as Video 045 Barge G4 Filmed Again After Repair Attempt

58 9/26/2008 Same as Video 047 Barge G4 Filmed Again After Repair Attempt

59 9/27/2008 Vent Barge G6

60 9/27/2008 Cofferdam Hatch Barge G6 Forward

61 9/27/2008 Cargo Hatch Barge G6 #3 Port



64 9/27/2008 Ullage Hatch Barge G6 #4 Starboard


66 9/27/2008 Same as Video 063 Barge G6 Filmed Again

67 9/27/2008 Same as Video 063 Barge G6 Filmed Again After Vent Was Closed

68 9/27/2008 Same as Video 063 Barge G6 Filmed Again After Vent Was Closed

69 9/28/2008 Overview of Leaks Barge G7 Overview of #2 & #3 Cargo Hatches

26


87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

9/28/2008

9/28/2008

9/28/2008

9/28/2008

9/28/2008

9/28/2008

9/28/2008

9/28/2008

9/28/2008

9/28/2008

9/28/2008

9/28/2008

9/28/2008

9/28/2008

9/28/2008

9/28/2008

9/28/2008

Pressure Relief Valve

Cargo Hatch

Cargo Hatch Control Valve

Control Valve Grease Cert

Hatch & Control Valve

Hatch & Control Valve

Overview of Leaks

Block Valve

Butterworth Hatch

Butterworth Hatch

Slop Tank Vent

Master Suction Valve

Butterworth Hatch

Cargo Hatch



Slop Tank Hatch

Barge G8

Barge G8 #2 Port

Barge G8 #2 Port

Barge G8 #2 Starboard


Barge G8 #3 Port

Barge G8 Overview of Videos 87 thru 92

Barge G8 #3

Barge G8 #3 Port Rear

Barge G8 #3 Starboard Rear

Barge G8

Barge G8

Barge G8 #2 Port Forward

Barge G8 #1 Port

Barge G8 #1 Port


Barge G8


September 2009

Figures 3-4 through 3-14 show snapshots of LSI PGIE observations onboard several barges acquired in conjunction with the LDEQ bagging. The figure captions also show the emission rate estimates from the LDEQ bagging survey report (Appendix H) converted to g/s for comparison purposes.

Note that it is impossible to represent the visual acuity of observed leaks with single image snapshots reproduced in this report. The leaks as viewed in moving video images are much more pronounced and generally easier to identify, particularly for small leaks.

27


September 2009

Figure 3-4. Example from LSI Ground Survey – Sampling During Bagging Test

Figure 3-5. Leak from Cargo Hatch on Barge G2- Mass Leak 1.86 g/s

28


September 2009

Figure 3-6. Leak from Ullage Hatch on Barge G4- Mass Leak 0.31 g/s


29


September 2009



30


September 2009


Figure 3-11. Leak from Vent on Barge G6- Mass Leak 1.45 g/s

31


September 2009

Figure 3-12. Leak from Cofferdam Hatch on Barge G6- Mass Leak 1.99 g/s

Figure 3-13. Leak from Cargo Hatch on Barge G7- Mass Leak 3.12 g/s

32


September 2009

Figure 3-14. Leak from Pressure Relief Valve on Barge G8- Mass Leak 5.78 g/s

Additional information on LSI ground videos can be found in Appendices C and D. Note that the screenshots shown in Appendix D do not include all detected leak events, but are only from events where the leaks are easily apparent in the screenshots taken from the camera videos.

The LSI ground surveys onboard the barges provided the following general conclusions about this remote sensing approach for fugitive emission detection from petrochemical transport barges:

1. The deployed PGIE equipment exhibited sufficient operational robustness and detection sensitivity and was judged to be extremely useful for type of close range fugitive leak inspection. This conclusion agrees with previous similar studies using this technology.

2. The PGIE technique was able to easily identify apparent large leaks with smaller leaks somewhat more difficult to identify, likely requiring an expert operator.

3. Ground surveys on board the barges indicated that the leaks originally seen from the air were many times composed of numerous individual leaks upon closer inspection.

33


September 2009

4. In all studied cases, leaks identified with the PGIE were verified as being significant by the LDEQ bagging study

5. Ground observations were easier to execute during mid-day to late afternoon time periods due to more favorable background imaging conditions (improved background radiance from hot barge surfaces and lower shadow interference) and because fugitive emissions were likely more pronounced as the barges became heated by solar radiation and ambient temperature during the day. Onboard observations were less sensitive to ambient conditions as compared to aerial observations (Section 3.1) due to the close-in nature of the inspection.

6. Onboard observations allow identification of much smaller leaks compared to airborne observations due to the close-in nature of the inspection and the ability to optimize viewing angles and focus.

7. Onboard PGIE observations are very useful for identification of specific leaking components and verification of subsequent leak repair activities.

3.3 PGIE Observations From the Lock Wall

Ground-based PGIE observations of fugitive emissions from petrochemical transport barges were made by LSI, LDEQ, and ARCADIS personnel from the Port Allen lock wall during the BEM 1 study. The purpose of the lock wall PGIE observations was to provide mid-range views of leaks from barges that were being measured by the OTM 10 OP-FTIR survey. PGIE Observations were conducted by LSI using the LSI FLIR camera on September 24 and September 28-30, 2008 and by LDEQ and ARCADIS using LDEQ FLIR camera on September 28 and October 1-9, 2008. The dataset contains multiple movies showing various leaks from several different barges. Table 3-3 lists the LSI observations with representative images contained in the Figures 3-15 through 3-17. Table 3-4 lists the LDEQ/ARCADIS PGIE observations from the lock wall with representative images contained in Figures 3-18 through 3-20. The tables contain video filename number, the date of the observation, and the barge number or other descriptor. Additional information and images are contained in Appendices C-E.

34

Table 3-3. Summary Table of Barge Leaks Identified by LSI Lock Wall Observations

Filename Date Part Leaking Barge # or Other Descriptor 0 1 2 3 4 5 6 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 104 105 106 107 108 109 110 111

9/24/2008 9/24/2008 9/24/2008 9/24/2008 9/24/2008 9/24/2008 9/24/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/28/2008 9/29/2008 9/29/2008 9/29/2008 9/29/2008 9/29/2008 9/29/2008 9/30/2008 9/30/2008

Hatch Hatch Hatch Hatch Hatch Hatch Hatch

Cargo Hatch Cargo Hatch Cargo Hatch Cargo Hatch

Same as Video 71 Pressure Relief Valve

Overview of Leaks Cargo Hatch

Alarm Test Rod Cargo Hatch Ullage Hatch Ullage Hatch Cargo Hatch Cargo Hatch

Vent Ullage Hatch Ullage Hatch Cargo Hatch Cargo Hatch

Hatch & Pressure Valve Butterworth & Cargo Butterworth Hatch

Cargo Hatch Hatch & Pressure Valve

Slop Tank Vent

Barge L1

Barge L1

Barge L1

Barge L1

Barge L1

Barge L2

Barge L2

Barge L3 #2 Port Barge L3 #2 Starboard Barge L3 #3 Starboard

Barge L3 #3 Port Barge L3 Filmed Again With Bag On

Barge L3 Barge L3 Another Overview of #2 & #3 Cargo Hatches

Barge L4 #3 Starboard Barge L4 #3 Port Barge L4 #2 Port Barge L4 #2 Port

Barge L4 #2 Starboard Barge L4 #1 Starboard

Barge L4 #1 Port Barge L4

Barge L4 #1 Port Barge L4 #1 Starboard

Barge L5 #1 Port Barge L5 #2 Port & #3 Starboard

Barge L6 #3 Port & Pressure Relief Valve Barge L6 #2 Starboard Forward & #1 Starboard

Barge L6 #1 Starboard Middle Barge L7 #3 Starboard

Barge L8 Pressure Relief Valve & #2 Starboard Barge L9


September 2009

35


September 2009

Figures 3-15 through 3-17 show three example snapshots from the lock wall observation illustrating the types of leaks that were detected: Example #1 shows a leaking hatch on Barge L1; Example #2 shows a leaking valve on Barge L2; and Example #3 shows a leaking hatch on Barge L6. The OTM 10 measured emission rates from barges in the lock are discussed in Sections 4 and 5.

Note that it is impossible to represent visual acuity of observed leaks with single image snapshots reproduced in this report. The leaks as viewed in moving video images are much more pronounced and generally easier to identify, particularly for small leaks.

Figure 3-15. Example #1 from LSI Ground Survey – Leaking Hatch on Barge L1- Total Mass Emission from Barge 0.521 g/s

36


September 2009

Figure 3-16. Example #2 from LSI Ground Survey – Leaking Valve on Barge L2- Total Mass Emission from Barge 0.521 g/s

Figure 3-17. Example #3 from LSI Ground Survey – Leaking Hatch on Barge L6- Total Mass Emission from Barge 0.415 g/s

37

Table 3-4. Summary Table of Barge Leaks Identified by LDEQ/ARCADIS Lock Wall Observations


VIDEO_080928_002

VIDEO_080928_002

VIDEO_081001_001

VIDEO_081001_002

VIDEO_081001_003

VIDEO_081001_004

VIDEO_081001_005

VIDEO_081001_006

VIDEO_081001_007

VIDEO_081002_001

VIDEO_081002_002

VIDEO_081002_003

VIDEO_081002_004

VIDEO_081002_005

VIDEO_081002_006

VIDEO_081002_007

VIDEO_081004_001

VIDEO_081004_002

VIDEO_081005_001

VIDEO_081005_002

VIDEO_081005_003

VIDEO_081006_001

VIDEO_081008_001

VIDEO_081008_002

VIDEO_081009_001

VIDEO_081009_002

9/28/2008

9/28/2008

10/1/2008

10/1/2008

10/1/2008

10/1/2008

10/1/2008

10/1/2008

10/1/2008

10/2/2008

10/2/2008

10/2/2008

10/2/2008

10/2/2008

10/2/2008

10/2/2008

10/4/2008

10/4/2008

10/5/2008

10/5/2008

10/5/2008

10/6/2008

10/8/2008

10/8/2008

10/9/2008

10/9/2008

Hatch

Hatch

Hatch

Hatch

Hatch

Hatch

Hatch

Valve

Vent

Hatch

Hatch

Hatch

Hatch

Hatch

Hatch

Hatch

Hatch

Hatch

Valve

Hatch

Hatch

Hatch

Vent

Vent

Vent

Vent

L10

L10

L11

L11

L11

L11

L11

L12

L12

L13

L13

L13

L13

L13

L14

L15

L16

L17

L18

L19

L20

L21

L22

L22

L23

L23


September 2009

38


September 2009

Figures 3-18 through 3-20 are three example snapshots illustrating the types of leaks that were detected: Example #1 shows a leaking hatch on Barge L10; Example #2 shows a leaking vent on Barge L23; and Example #3 shows a leaking valve on Barge L18. Additional example screenshots from the LDEQ camera videos can be found in Appendix E with discussion on OTM 10 emission flux measurement in Sections 4 and 5.

Figure 3-18. Example #1 from Ground Survey with LDEQ Camera – Leaking Hatch from Barge L10 (there was no OTM 10 emission flux measurement for this time period)

39


September 2009

Figure 3-19. Example #2 from Ground Survey with LDEQ Camera – Leaking Vent from Barge L23- Total OTM 10 Mass Emission from Barge 0.490 g/s

Figure 3-20. Example #3 from Ground Survey with LDEQ Camera – Leaking Valve from Barge L13- Total OTM 10 Mass Emission from Barge 0.106 g/s

40


September 2009

The lock wall PGIE observations provided the following general conclusions about this remote sensing approach for fugitive emission detection from petrochemical transport barges:

1. The deployed PGIE equipment exhibited sufficient operational robustness and detection sensitivity and was judged to be extremely useful for this type of mid-range distance fugitive leak inspection. This conclusion agrees with previous similar studies using this technology.

2. The PGIE technique was able to easily identify apparent large leaks with smaller leaks somewhat more difficult to identify, likely requiring an expert operator.

3. Lock wall mid-range observations were easier to execute during mid-day to late afternoon time periods due to more favorable background imaging conditions (improved background radiance from hot barge surfaces and lower shadow interference) and because fugitive emissions were likely more pronounced as the barges became heated by solar radiation and ambient temperature during the day. Lock wall observations were affected by strong shadows present in the deep lock under certain lighting conditions.

4. Mid-range Lock wall PGIE observations allow identification of much smaller leaks compared to airborne observations but were not as sensitive as close-range inspection onboard the barges which benefited from shorter range and the ability to optimize viewing angles.

5. Lock wall PGIE observations are judged to be very useful for routine inspection of barges passing through the lock.

41


September 2009

4. OTM 10 AM Emission Flux and Trace Compound Speciation Results

The OTM 10 measurement at the Port Allen lock was performed from September 24 through October 9, 2008 from approximately 8:00 a.m. to 5:00 p.m. each day. The OTM 10 measurement attempted to assess emissions from all barge traffic that passed through the lock during this time period. In addition to OTM 10 measurements, PGIE camera images were acquired to compare measured flux rates leak appearance (Section 3.3). OTM 10 measurements focused on quantification of an alkane mixture (AM) flux (Appendix F).

Recorded barge traffic data included time of lock entry and exit and visual inspection of cargo labels on each barge. The reported cargo from the U.S. Army Corps of Engineers traffic log was also recorded. Note that the Corps of Engineers lock staff advised that the lock traffic reports are not necessarily accurate with regard to barge cargo. Of the six highest emissions events recorded by OTM 10, two occurred during times with barges that coded as carrying petroleum pitches, two with barges coded as carrying crude petroleum, and two with barges coded as empty (however the field crew smelled aromatics during one of these events). All OTM measured events and barge information are contained in Appendix G with a subset of the most interesting events presented in this section

4.1 Data Graphs and Tables for Select Events

A total of 97 barge sets (one or more barges per tug) passed through the lock during the OTM 10 observation period. AM fluxes were measured in a total of 62 temporally defined events. Many of these events exhibited small but measureable AM fluxes (< 0.1 g/s) and occurred when non-petrochemical transport barges were in the lock indicating that the measured AM emissions were possibly associated with hydrocarbon emissions from the tug diesel engines from the tugs idling in the lock. A significant portion of the events exhibited high AM flux emissions and occurred in conjunction with PGIE leak identification from lock wall observations. These events are believed to be related to fugitive emissions from the barge with only a small relative component of AM emissions from tugs.

A subset of OTM 10 measurements is provided below with a complete listing contained in Appendix G. The summary contains a description of the event, the AM flux values measured during the event (presented in Tables 4-1 through 4-12), and a screenshot of a leak detected during the event from the PGIE observations when available (presented in Figures 4-1 through 4-6). The spectral data from each of these events were also screened for trace VOC compounds. The results of the trace compound analysis (when detected) are presented in Section 4.1.1. For some of the events, we report “WC” as the AM flux value. In these instances, AM concentrations were detected by the OP-FTIR instrumentation, but the

42

Table 4-1. 9/24/ 2008 ─ Event #1

Date Entry Time Exit Time Number of Barges Description of Commodity

9/24/2008 10:32 11:03 Two Labeled as benzene. Reported as empty. Noticeable aromatic smell,


September 2009

prevailing winds during the time of the measurement did not meet minimum data quality indicator levels regarding normal wind direction so an AM flux value could not be calculated.

Table 4-2. AM Flux Values Measured during 9/24/ 2008, Event #1

Time AM Flux (g/s)

10:36:59 0.124

10:39:37 0.237

10:42:17 0.321

10:44:56 0.431

10:47:35 0.558

10:50:13 0.637

10:52:53 0.730

10:55:33 0.912

10:58:12 0.956

11:00:15 0.308

Average: 0.521

43


September 2009

Figure 4-1. Screenshot from FLIR Camera Showing Leak from 9/24/ 2008, Event #1

Table 4-3. 9/29/ 2008 ─ Event #2


Three tugs with 9/29/2008 9:23 10:23 barges, one empty Equipment/machinery/other

and two manned

44

Table 4-4.

Time

9:39:26

AM Flux Values Measured during 9/29/ 2008, Event #2

AM Flux (g/s)

0.514

9:42:06 0.374

9:44:45 0.305

9:47:23 0.325

9:50:00 0.410

9:52:39 0.457

9:55:16 0.596

9:57:55 0.590

10:00:33 0.463

10:03:11 0.324

10:05:50 0.297

10:08:29 0.389

10:11:09 0.374

10:13:48 0.390

Average:

0.415


September 2009

45

Table 4-5. 10/1/ 2008─ Event #2


10/1/2008 9:10 9:58 Two Organic industrial chemicals


September 2009


46


September 2009


Time AM Flux (g/s)

9:14:56 wc

9:17:33 wc

9:20:12 wc

9:22:51 wc

9:25:30 wc

9:28:09 0.019

9:30:52 wsc

9:33:30 0.035

9:36:09 0.046

9:38:54 0.067

9:41:31 0.065

9:44:12 0.058

9:46:48 0.04

9:49:23 wc

9:52:01 wc

Average: 0.047

wc = Wind criteria were not met.

wsc = Wind speed criteria not met.

47


September 2009


Table 4-7. 10/2/ 2008 ─ Event #2


10/2/2008 9:45 10:42 One tug with one barge, one tug with two barges Butane, propylene, one empty

48


September 2009


Time AM Flux (g/s)

9:48:36 wc

9:51:16 wc

9:53:53 wc

9:56:31 wc

9:59:09 0.072

10:01:46 0.141

10:04:24 0.126

10:07:02 0.084

10:09:41 wc

10:12:18 wc

10:14:57 wc

10:17:35 wc

10:20:11 wc

10:22:48 wc

10:25:28 wc

10:28:05 wc

10:30:41 wc

10:33:21 wc

10:36:00 wc

10:38:38 wc

Average: 0.106

wc = Wind criteria were not met.

49

Table 4-9. 10/5/ 2008 ─ Event #1


10/5/2008 9:23 10:04 Two Petroleum


September 2009


50

Table 4-10.

Time

9:26:53


AM Flux (g/s)

2.48

9:29:31 4.07

9:32:09 4.84

9:34:48 4.02

9:37:25 4.03

9:41:21 4.31

9:43:59 3.30

9:46:40 3.00

9:49:18 2.85

9:51:56 2.90

9:54:35 2.33

9:57:13 2.57

Average:

3.39


September 2009

51

Table 4-11. 10/9/2008- Event #7


10/9/2008 14:47 15:25 Two Petroleum products, Empty


September 2009


52

Table 4-12.

Time

14:51:46


AM Flux (g/s)

0.286

14:54:24 0.331

14:57:02 0.635

14:59:40 0.794

15:02:17 0.819

15:04:55 0.877

15:07:32 0.661

15:10:11 0.432

15:12:49 0.206

15:15:28 0.18

15:18:05 0.174

Average:

0.490


September 2009


53


September 2009

4.1.1 Summary of the Results of Analysis of Trace VOC Concentrations

As mentioned above, the spectral data from each of the six events described above were screened for the presence of trace VOC. The data were collected with the EPA OP-FTIR along an optical beam path that extended along the surface of the lock, from one side of the lock to the other. The results of this analysis are presented in Table 4-13. The table also includes the measured alkane mixture concentration for each event. The instrument minimum detection limit for each compound is shown in parentheses. Additional VOC analysis was performed for all other temporally defined events during the OTM 10 measurements. The results of this analysis are presented in Appendix G of this document.

Table 4-13. Summary of VOC Analysis

Event Date

Alkane Mixture (ppb)

Mol. Mass Alkane Mixture

(g/mole) Benzene

(ppb) Toluene

(ppb) m-Xylene

(ppb) Styrene

(ppb) Ethylene

(ppb) 1,3-Butadiene

(ppb) Methane*

(ppb)

9/24/08 1002 79 ND(47) ND(77) 37(34) ND(11) ND(7) ND(12) 115

9/29/08 736 61 ND(61) ND(96) ND(52) ND(17) ND(9) ND(18) ND

10/1/08 184 58 ND(41) ND(51) ND(26) ND(16) 8(7) 17(13) 302

10/2/08 1954 68 ND(44) 73(51) ND(27) 18(12) ND(7) ND(10) 114

10/5/08 3826 62 71(66) ND(80) ND(36) ND(21) ND(10) 16(15) 172

10/9/08 768 60 ND(55) ND(85) ND(39) ND(13) ND(7) ND(18) 40

*Methane concentrations reported were measured along the ground-level beam path of the EPA OP- FTIR OTM 10 Configuration

ND = Not detected

4.2 Instances of Emissions Detected with the PGIE but not with ORS Measurements

An analysis of the PGIE observations made by the LSI Ground Crew and ARCADIS personnel in the lock revealed that there were instances where the PGIE detected barge leaks, but the events were not detected by the ORS instrumentation deployed on the southern side of the lock. Table 4-14 presents a summary of six events that were detected by the PGIE but not the ORS instrumentation. The table includes the date and time of the events, as well as the average prevailing wind direction during the time the PGIE detected the leaks.

54

Table 4-14. Summary of Leak Events Detected by the PGIE but not ORS Instrumentation

Date Time Barge Number(s) Prevailing Wind Direction (degrees)

9/28/08 11:30 am L10 120

9/29/08 4:32 pm L7 320

9/30/08 2:46 pm L8 300

10/2/08 10:25 am L18, L19 140

10/2/08 1:00 pm L14 180

10/2/08 2:35 pm L15 150

10/8/08 9:21 am L22 320


September 2009

The orientation of the ORS measurement planes (when looking from the OP-FTIR to the scissor lift) were 133° and 311° for the EPA and ARCADIS OP-FTIR measurement planes, respectively. Considering the ORS configurations used in the study, a prevailing wind direction of approximately 41° is ideal for emissions measurements (perpendicular to the configuration planes). As can be seen in Table 4-14, the prevailing winds during the events not detected by the ORS instrumentation were close to parallel to the measurement configurations, or in some cases the winds were not from the direction of the lock (wind direction greater than 133° or less than 311°). The prevailing winds during the times the leaks were detected by the PGIE did not carry the winds through the ORS measurement plane, which explains why the leaks were not detected by the ORS instrumentation.

4.3 Evaluation of AM Emissions from Tugs

In order to evaluate the contribution of exhaust from the tugs to the alkane mixture (AM) emissions fluxes measured during the project, carbon monoxide concentrations were analyzed along the ground level beam path of the ARCADIS OP-FTIR VRPM configuration. Carbon monoxide was chosen for this analysis because it is a byproduct of combustion, and has relatively low detection limits with the OP-FTIR instrument. For the nine events detected from barges classified as “empty-no further information”, the carbon monoxide and alkane mixture concentrations measured along the ground level beam path were compared to investigate any possible correlations between the two compounds. A correlation between the two compounds would suggest that the source of the total hydrocarbon emissions measured was the emissions from the tug engines.

Of the nine events analyzed, eight of the events showed no correlation between the measured carbon monoxide and total hydrocarbon concentrations. The analysis did indicate a strong correlation between the concentrations of the two compounds during the 9/28/08

55


September 2009

9:38 am to 10:11 am event (r2 =0.83). However, the alkane mixture concentrations measured during this event were relatively low and close to the minimum detection limits of the OPFTIR instrument. Based on these findings, we conclude that emissions of alkane mixture from the tug exhaust are negligible. The data from this analysis are presented in Appendix I of this document.

The OTM 10 lock wall mass flux measurements provided the following general conclusions about this remote sensing approach for fugitive emission detection and quantification from petrochemical transport barges:

(1) The deployed OTM 10 equipment exhibited sufficient operational robustness and detection sensitivity and was judged to be useful for this type of mid-range distance leak detection/quantification activities where compound speciation is important.

(2) The OTM 10 technique was able to identify and assess emission rates from a range of leak sizes as long as the prevailing wind brought the emitted plume through the vertical plane of the EPA OTM-10 measurement configuration.

56


September 2009

5. Bagging Test Emission Estimate Results

As described previously, onboard leak bagging tests were performed by SAGE Environmental Consulting for LDEQ (Appendix H). The testing was performed directly at the source of the leak on each barge. During the five day bagging test, 23 leak points from a total of eight barges were measured to determine approximate THC mass emission rates. Table 5-1 reproduces the bagging test results contained in the LDEQ report including the measured total non-methane hydrocarbon emissions with values converted to g/s for comparison. The table shows that the measured total non-methane hydrocarbon emissions flux values ranged from 0.07 to 5.77 g/s.

Table 5-1. Summary of LDEQ Bagging Test Results

Test# Barge# Cargo Mass Leak (g/s)

1 G1 Unleaded Gasoline 2.53 2 G2 Trans Mix 0.31 3 G2 Trans Mix 0.57 4 G2 Trans Mix 1.86 5 G2 Trans Mix 0.32 6 G3 Trans Mix 0.89 7 G3 Trans Mix 1.32 8 G4 Naphtha but cleaned 0.31 9 G4 Naphtha but cleaned 0.18 10 G4 Naphtha but cleaned 0.24 11 G4 Naphtha but cleaned 0.13 12 G4 Naphtha but cleaned 0.07 13 G4 Naphtha but cleaned 0.20 14 G5 Raffinate 2.11 15 G5 Raffinate 0.73 16 G5 Raffinate 1.42 17 G5 Raffinate 0.07 18 G6 Gasoline 1.45 19 G6 Gasoline 1.98 20 G7 Naphtha 3.12 21 G7 Naphtha 0.66 22 G8 Unleaded Gasoline 5.77 23 G8 Unleaded Gasoline 0.47

57


September 2009

6. Comparison of OP-FTIR and Bagging Test Emission Flux Results

It is instructive to compare OTM 10 measurements and LDEQ bagging study results to help draw overall conclusions regarding the study. However, it is important to note that the results of the bagging method report emissions flux values from each individual leak, while the OTM 10 results report emissions flux values for each barge (possible multiple barges) and could consist of multiple leaks. Additionally, as mentioned in a previous section of the report, barge traffic through the Port Allen Lock was much lighter than normal due to repair activities on the Intracoastal Waterway Bayou Sorrel Bridge. Although study personnel originally planned to conduct bagging tests on barges immediately upon exiting the lock, this was not possible due to a lack of barge traffic through the lock during the period that Sage Environmental personnel were at the site. Instead, the barges selected for monitoring using the bagging method were chosen because they were identified as having very significant leaks from the airborne PGIE surveys. Therefore, the barges selected for the bagging experiments may not represent an average case, whereas the OTM 10 measurements were conducted on barges moving through the lock with no selection process and may represent a more typical sample cross-section. Even with these differences, it is useful to compare the results of emissions flux values determined from each method on different barges.

As discussed, the LDEQ bagging test results report findings for individual leaks on the barge as compared to the OTM 10 results which can include emissions from multiple leak points on a given barge or multiple barges. Table 6-1 presents results from the two measurement methods expressing the results of the LDEQ bagging test as a summation of measured leaks from a given barge and tabulating this with the most significant OTM 10 flux rate measurements. Note the measurements are from different barges so they serve as only a range comparison. The LDEQ bagging test results are labeled (Bag) and show leak rates of THC whereas the OTM 10 results, labeled (OTM 10), show AM flux rates.

58

Table 6-1.

Test

Bag

Bag

Bag

Bag

Summary of LDEQ

Barge Number(s)

G1

G2

G3

G4

Bagging Test Barge Totals and Mo

Cargo

Unleaded Gasoline

Trans Mix

Trans Mix

Naphtha but cleaned

st Significant OTM 10 Results

Total Mass Leak Rate

(g/s)

2.53

3.06

2.21

1.13

Bag

Bag

Bag

Bag

G5

G6

G7

G8

Raffinate

Gasoline

Naphtha

Unleaded Gasoline

4.33

3.43

3.78

6.24

OTM 10

OTM 10

OTM 10

OTM 10

L1

L5, L6

L11

L13

Benzene (Empty)

Equipment/Machinery/Other

Organic industrial chemicals

Butane, propylene, one empty

0.521

0.415

0.047

0.106

OTM 10 L18, L19 Petroleum 3.39

OTM 10

L23 Petroleum products, one empty 0.490


September 2009

The above comparison is for different barges using different measurement techniques each of which can possess significant uncertainty due to the difficulty of the assessment. The primary sources of uncertainty are described in Section 7. With the difficulties these measurements present, the relative agreement between the two techniques may provide some confidence in the individual measures. From Table 6-1 the range of AM flux values found with the OTM 10 method was generally lower than the values found using the bagging method although the maximum flux values measured are comparable (3.39 with the OTM 10 method and 6.24 with the bagging method). The barges selected for the bagging experiments were identified as having very significant leaks from the airborne survey so they may not represent an average case whereas the OTM 10 measurements were conducted on barges moving through the lock with no selection process and therefore represent a more typical sample cross-section. This fact could help explain the lower values observed by OTM 10. Additionally, we would expect some underestimation of the alkane mixture (AM) flux measurement by OP-FTIR in comparison to the total non-methane hydrocarbon measurement produced in the bagging studies since non-alkane compounds can be somewhat underrepresented in the OP-FTIR AM approximation due to lack of signal in the spectral analysis region.

59


September 2009

Note that these emission estimates presented in this report represent a snapshot in time. Fugitive emissions from petrochemical barges are believed to vary significantly due to ambient temperature, thermal load, product mix, load state, and equipment condition.

60


September 2009

7. Quality Assurance/Quality Control

The following sections discuss the quality assurance methods used for the OTM 10 measurements. Note that quality assurance methods and procedures for the bagging tests can be found in Appendix H of this document.

7.1 Instrument Calibration

As stated in the ECPD Optical Remote Sensing Facility Manual (U.S. EPA, 2004), all equipment is calibrated annually and / or cal-checked at the U.S. EPA facility as part of standard operating procedures. Certificates of calibration are kept on file. Maintenance records are kept for any equipment adjustments or repairs in bound project notebooks that include the data and description of maintenance performed. Instrument calibration procedures and frequency are listed in Table 7-1 and further described in the text.

Table 7-1. Instrumentation Calibration Frequency and Description

Instrument Measurement Calibration Date Calibration Detail

IMACC, Inc. OP-FTIR Analyte PIC Pre-deployment and in-field checks

MOP-6802 and 6823 of the ECPB Optical Remote Sensing Facility Manual

R.M. Young Model 05103 Meteorological Head

Wind Speed in miles/hour 6/21/08 U.S. EPA Wind Tunnel Cal. Records on

file at EPA Metrology Lab

R.M. Young Model 05103 Meteorological Head

Wind direction in degrees from North 6/21/08 U.S. EPA Wind Tunnel Cal. Records on

file at EPA Metrology Lab

Topcon Model GTS-211D Theodolite

Distance Measurement 6/17/08

Calibration of distance measurement.

Actual distance = 42.5 ft

Measured distance = 43.11 ft

Topcon Model GTS-211D Theodolite Angle Measurement 6/17/08

Calibration of angle measurement.

Actual angle = 360º

Measured angle = 359º01’08”

7.2 Assessment of DQI Goals

The critical measurements associated with this project and the established data quality indicator (DQI) goals in terms of accuracy, precision, and completeness are listed in Table 7-2. More information on the procedures used to assess DQI goals can be found in Section 10 of the ECPD Optical Remote Sensing Facility Manual (U.S. EPA, 2004).

61


September 2009

Table 7-2. Data Quality Indicator Goals for the Project

Measurement Parameter Analysis Method Accuracy Precision Completeness

Analyte PIC OP-FTIR: Nitrous Oxide Concentrations ± 25%/15%/10% a ± 10% 90%

Ambient Wind Speed

R.M. Young Met heads pre-deployment calibration in

EPA Metrology Lab ± 1 m/s ± 1 m/s 90%

Ambient Wind Direction

R.M. Young Met heads pre-deployment calibration in

EPA Metrology Lab ± 10º ± 10º 90%

Distance Measurement Theodolite- Topcon ± 1m ± 1m 100%

Gas plume relative opacity PGIE: gasoline vapor release N/Ab N/Ab 100%

(a) The accuracy acceptance criterion of ± 25% is for path lengths of less than 50 m, ± 15% is for path lengths between 50 and 100 m, and ± 10% is for path lengths greater than 100 m.

(b) The PGIE is not a quantitative device and does not provide a numerical output.

7.2.1 DQI Check for Analyte PIC (OP-FTIR) Measurement

The precision and accuracy of the concentration data may be checked by looking at the analyzed nitrous oxide concentrations. The known atmospheric background nitrous oxide concentration is around 315 ppbv (this is an average value, as the value exhibits a slight seasonal variation). The acceptable range of nitrous oxide concentrations is 315 ppb ± 25% for path lengths of less than 50 m, 315 ppb ±15% for path lengths between 50 and 100 m, and 315 ppb ±10% for path lengths greater than 100 m. Verifying this background concentration provides a good QC check of the data collected. Obviously, this method is not valid for data collected at a site that is a source of nitrous oxide.

The precision of the analyte PIC measurements was evaluated by calculating the relative standard deviation (RSD) from one data subset collected near the surface of the suspected source. A subset is defined as the data collected along one particular path length during one particular survey in one survey sub-area.

The accuracy of the analyte PIC measurements was evaluated by comparing the calculated nitrous oxide concentrations from one data subset to the background value of 315 ppb. The number of calculated nitrous oxide concentrations that failed to meet the DQI accuracy criterion was recorded.

62


September 2009

Overall, a total of three datasets were analyzed from three different time periods, one at the beginning of the project (September 24), one during the middle of the project (September 28), and one at the end of the project (October 9). Based on the DQI criterion set forth for precision of ±10%, all of the data subsets were found to be acceptable, for a completeness of 100%. The range of calculated relative standard deviations for the data subsets from this field campaign was 1.8 to 4.0 ppb, which represents 0.57 to 1.27% RSD.

Each data point (calculated nitrous oxide concentration) in the data subsets was analyzed to assess whether or not it met the DQI criterion for accuracy of ±10% (315 ± 32 ppb), as the path lengths used for measurements were greater than 100 meters. A total of 233 data points were analyzed, and 176 of the points met the DQI criteria for accuracy for a completeness of 76%.

7.2.2 PGIE Relative Opacity DQI Assessment

The PGIE used in this study are not quantitative instruments and therefore do not provide calibrated numerical data for images. The performance of the device can be assessed in a basic way by imaging a known gas release against a stable background, such as a concrete pad. During the current campaign, the vapors from an opened container of gasoline were used for the known gas release. Imaging this test release ensured the camera was operating properly and device firmware was set correctly.

7.2.3 Meteorological Head DQI Assessment

The meteorological head DQIs are checked annually as part of the routine annual calibration procedure. Before deployment to the field, the user verified the calibration date of the instrument by referencing the calibration sticker. If the date indicates the instrument is in need of calibration, it should be returned to the APPCD Metrology Laboratory before deployment to the field. The precision and accuracy of the heads is assessed by conducting a calibration in the EPA Metrology Lab using the exhaust from a bench top wind tunnel. This calibration procedure differs from the procedure used to perform the annual calibration of the instruments.

Additionally, a couple of reasonableness checks were performed in the field on the measured wind direction data. While data collection is occurring, the field team leader compares wind direction measured with the heads to the forecasted wind direction for that particular day. Another reasonableness check involves manually setting the vane on the meteorological heads to magnetic north (this is done with a hand held compass). The observed wind direction during this test should be very close to 360º.

63


September 2009

7.2.4 Topcon Theodolite DQI Assessment

Before field deployment, ensure the battery packs are charged for this equipment. The following additional checks (which are performed at least annually) were made on June 17, 2008. The calibration of distance measurement was done at the EPA Facility using a tape measure. The actual distance was 42.5 feet, and the measured distance was 43.11 feet. The results indicate that instrument accuracy falls well within the DQI goals. The calibration of angle measurement was also performed. The actual angle was 360°, and the measured angle was 359°01’08”. The results indicate accuracy falls well within the DQI goals.

Additionally, there are several internal checks in the theodolite software that prevent data collection from occurring if the instrument is not properly aligned on the object being measured, or if the instrument has not been balanced correctly. When this occurs, it is necessary to re-initialize the instrument to collect data.

7.2.5 QC Checks of OP-FTIR Instrument Performance

At the beginning of the project, a series of QC checks were performed on both OP-FTIR instruments to assess the instrument performance. On September 25, 2008, the Single Beam Ratio, Baseline Stability, Noise Equivalent Absorbance, ZPD Stability, Saturation, Random Baseline Noise, and Stray Light diagnostic tests were performed. The results of the tests indicated that the ARCADIS and EPA OP-FTIR instruments were operating within the acceptable criteria for each QC check. More information on the diagnostic checks that are performed as part of a typical ORS field campaign can be found in MOP 6802 and 6823 of the ECPD Optical Remote Sensing Facility Manual (U.S. EPA, 2004).

In addition to the QC checks performed on the OP-FTIR instruments, the quality of the instrument signals (interferogram) was checked constantly during the field campaigns by ensuring that the intensity of the signal was at least 5 times the intensity of the stray light signal (the stray light signal is collected as background data prior to actual data collection, and measures internal stray light from the instrument itself). In addition to checking the strength of the signal, checks were done constantly in the field to ensure that the data were being collected and stored to the data collection computer. During the campaign, a member of the field team monitored the data collection computer to make sure these checks were completed.

64


September 2009

7.3 Estimate of Uncertainty in the OTM 10 Emission Flux Measurements

As mentioned in Section 2, the OTM 10 measurement configurations consisted of three measurement paths which extended from the OP-FTIR instrument to the scissor lift. The 3beam OTM 10 approach was chosen for this project since it was much more important to obtain a larger number of measurement cycles while the mobile source was contained in the lock rather than a fewer number of cycles with a five beam approach since the horizontal spatial location of the plume was not of primary importance. It was generally assumed that the plumes emitted from the barges would be initially small in spatial extent but would likely experience significant dispersion before exiting the lock and passing through the OTM 10 plane. It is likely that this assumption is correct since the barges were significantly below the lock wall top (approximately 7m to 12 m) and the emitted plumes could experience several dispersive/mixing mechanisms (such as stagnation, turbulence, channeling) depending on ambient wind direction and speed so the plumes could evolve more than in a flat wind swept scenario with similar downwind standoff.

In analyzing the PIC data using the 3-beam approach, the peak plume concentration was assumed to be centered along the crosswind axis of the OTM 10 configuration, and the σy

parameter (horizontal dispersion coefficient) of the measured plume was assumed to be equal to ½ the length of the OTM 10 configurations. It was necessary to make these assumptions because the 3-beam OTM 10 approach does not include two intermediate surface beam paths which are used to obtain information on the horizontal location and dispersion of the plume.

In order to estimate the uncertainty associated with assuming a fixed peak plume concentration location and σy parameter, we used the VRPM Fit Explorer program (described by Abichou et al., 2009) to run a series of simulations to assess the variability in flux results from the OTM 10 method as a result of assuming different σy and peak plume concentration locations. In this simulation program, a downwind concentration field is generated from an area source using EPA ISC Gaussian dispersion model and then analyzed using OTM 10 algorithms and optical beam geometries.

Table 7-3 presents the results of a simulation done for three different assumed plume sizes where the σy parameter was varied, but the plume location was assumed to be fixed in the center of a 160 meter measurement configuration. The plume dimensions are shown in Table 7-3 as: width (m) by crosswind distance (m). The results are shown in units of g/s, and are compared to an OTM 10 calculation of 1.0 g/s for the same plume size assuming a σy

value of 80 meters and a peak plume concentration location at 80 meters (as measured

65


September 2009

Peak Plume Concentration Location 5 m × 40 m 5 m × 80 m 5 m × 120 m

(m)

20 3.651 N/A N/A

40 2.066 1.910 N/A

60 1.305 1.322 1.300

80 0.921 0.945 0.978

100 0.700 0.718 0.743

120 0.545 0.554 N/A

140 0.412 N/A N/A

along the surface of the OTM 10 configuration plane from the OP-FTIR instrument), which is approximately equal to the parameters used for data analysis in the current study.

Table 7-3. Results of Flux Values Calculated by the VRPM Fit Explorer Program With a Fixed Peak Plume Concentration Location and Varying Values of the σy Parameter

σy Value 5 m × 40 m 5 m × 80 m 5 m × 120 m

σy = 8 m 1.002 1.030 1.068

σy = 80 m 0.921 0.945 0.978

σy = 800 m 0.914 0.938 0.970

The results of the simulation show that the OTM 10 calculation is insensitive to varying the value of the σy parameter (the OTM 10-derived flux values from the simulation were within ± 8.6% of control simulated values).

Table 7-4 presents the results of a second simulation done for three different assumed plume sizes where the plume center location was varied, but the σy parameter was assumed to be 80 m. The results are shown in units of g/s, and are compared to an OTM 10 calculation of 1.0 g/s for the same plume size assuming a σy value of 80 meters and a peak plume concentration location at 80 meters.

Table 7-4. Results of Flux Values Calculated by the VRPM Fit Explorer Program with a Fixed σy Parameter and Varying Peak Plume Concentration Locations

N/A- Simulation results not included because plume would not be located within the confines of the OTM 10 configuration plane

66


September 2009

The results of the simulation show that the 3-beam OTM 10 calculation is highly dependent upon the peak plume concentration location along the OTM 10 configuration plane. When the simulation was run with the peak concentration location close to the location of the OPFTIR instrument (peak concentration location < 40 m), the OTM 10-derived flux values from the simulation were as much as 265% higher than control simulated values. However, the OTM 10-derived flux values from the simulation agreed better with control values as the plume becomes larger and is more centered on the optical configuration. Underestimation is evident closer to the end of the configuration defined by the location of the scissor lift.

Additional analysis was performed to compare the AM concentrations measured on the lowest beam path of each OTM 10 configuration during the AM emissions events described in Section 4.1. The analysis showed the average AM concentration ratio of the lowest OTM 10 beam paths for the two measurement planes for the several measurement periods presented in the report (in ppb) are: 324/63, 271/59, 244/39, 1563/945, 306/211, with baseline levels below 10 ppb. This suggests that the peak plume concentration location for each emissions event was located at a position along the OTM 10 configuration plane closer to the scissor lift (≥ 80 m) and/or that the effective plume size was more similar to the 120m (large plume) case. Based on the above plane to plane ratio analysis, a small plume located near the vertex of the beams (highly overestimated case) was not likely. Based on the simulation results and information on OTM 10 measurement accuracy from previous tracer-release validations studies, it is reasonable to assume that the overall uncertainty in the AM flux results for this effort are likely within ± 50%.

7.4 Uncertainty in the LDEQ Leak Bagging Estimates

The on-board leak bagging measurements conducted by SAGE Environmental Consultants for LDEQ is described in Appendix H of this report. Sage identifies several factors which can impact uncertainty in mass emission estimates including: sampling and analytical variability, leak capture/containment variability, inter-dependence of multiple leaks, and temperature effects. Additionally, this measurement campaign represented the first attempt, to our knowledge, to produce leak emission rate estimates from these source types using the component bagging technique. As a consequence, there is inherent uncertainty associated with novel application. To supplement Appendix H, further information on the execution of the bagging study and potential areas leading to uncertainty can be found in Appendix J which reproduces comments from the American Waterways Operators on this testing procedure along with responses from Sage Environmental Consulting.

67


September 2009

7.5 General Data Limitations

One aspect of this report centers on the use of optical remote sensing equipment (especially PGIE) for identification of significant fugitive leaks from this difficult to measure source category in real-world scenarios. There are few perceived limitations on the use of the acquired remote sensing and imagery data in support the conclusion that these tools are generally useful for this purpose. Questions still remain as to the limits of detection and how these limits are affected by various field, target, and instrument parameters and these questions should be the subject of further study.

Another aspect of this report relates to estimates of emissions from this source category. Both measurement techniques, (EPA OTM 10 from the lock wall and the LDEQ bagging study) produced results which indicate a potentially significant level of short-term fugitive VOC emissions can occur. As discussed, measurements from this source category are difficult, and there is significant uncertainty in the absolute measurement results from both techniques and this should be considered a limitation of the data. Additionally, a more significant data limitation centers on the short-duration nature of these measurements which represent a snap-shot in time. Fugitive emissions from petrochemical barges are believed to vary significantly due to ambient temperature, thermal load, product mix, load state, and equipment condition. Since there is little information on the influence of these factors, extrapolation of these short term emission rate estimates is not recommended.

7.6 Deviations from the QAPP

The Quality Assurance Project Plan indicated that an ultraviolet differential optical absorption spectroscopy (UV-DOAS) instrument would be deployed to collect supplemental measurements of the BTEX compounds. The UV-DOAS instrument was not deployed at the site due to limited project resources and potential eye safety issues at the site. Instead, two additional OP-FTIR optical beam paths were deployed (one from each OP-FTIR instrument) across the surface of the lock to collect supplemental data on alkane mixture and trace VOC concentrations.

Also, it was originally anticipated that the VRPM configuration would be deployed along the northern edge of the lock. However, at the time of the field campaign, the winds were largely from the north, and the configuration was deployed on the southern edge of the lock.

68


September 2009

8. Summary

This report describes the BEM 1 field campaign conducted in Baton Rouge, Louisiana from September 24 to October 9, 2008. BEM 1 investigated VOC emissions from petrochemical transport barges using portable gas imaging equipment PGIE (infrared cameras), and EPA Method OTM 10 with Open-path Fourier transform infrared (OP-FTIR) spectrometers, in addition to leak bagging tests.

The objectives of the study were:

• To improve knowledge of fugitive VOC emissions from petrochemical transport barges.

• To demonstrate and advance the field application of select ORS techniques (EPA OTM 10 OP-FTIR and PGIE) for identification and quantification of fugitive emissions from difficult to monitor sources.

• Identify sources of fugitive leaks from multiple barges

To accomplish these goals, the project team conducted several complementary efforts:

1. Aerial PGIE surveys of barges located on the Mississippi River and inter-coastal water ways to identify barges with significant fugitive emissions.

2. Ground-based PGIE observations of barges from the Port Allen Lock wall and also onboard several barges to identify and closely observe fugitive leaks.

3. Onboard leak emission bagging measurements conducted by LDEQ on several barges to quantify leak rates and allow comparison with PGIE images.

4. EPA method OTM 10 with open-path Fourier transform infrared spectroscopy used at the Port Allen lock to produce hydrocarbon emission measurements from barge traffic traveling through the lock.

The aerial PGIE camera monitoring performed by LSI, Inc. detected leaks from 45 different barges located in the Mississippi River and the Intracoastal Waterway. The ground-based monitoring performed by LSI, Inc. detected leaks from 18 different barges in the U.S. Army Corps of Engineers lock and in the Mississippi River. Additional infrared camera monitoring performed by ARCADIS and LDEQ personnel in the lock detected multiple leaks from

69


September 2009

several barges. This report contains a number of leak images that serve to further understanding of fugitive emissions from various barge components.

The remote sensing surveys provided significant information regarding the use of infrared cameras for detection of fugitive emissions from petrochemical transport barges. The PGIE equipment was robust, easy to use, and possessed sufficient detection sensitivity for this application. The PGIE remote sensing approach is judged to be extremely useful for both aerial survey and close range fugitive leak inspection of petrochemical transport barges. The PGIE technique was able to identify a large range of leaks with large leaks detectable from the air and smaller leaks more easily observed at close range. PGIE observations were easier to execute during mid-day to late afternoon time periods due to more favorable background imaging conditions (improved background radiance from hot barge surfaces and lower shadow interference) and because fugitive emissions were likely more pronounced as the barges became heated by solar radiation and ambient temperature during the day. PGIE observations were very useful for identification of specific leaking components and verification of subsequent leak repair activities.

Based on aerial observations, eight barges with observed large leaks were selected for onboard leak emission rate measurements as part of the LDEQ on-board bagging survey. For this effort, a total of 23 leak points from eight barges were bagged to determine mass emission rates. The measured total non-methane hydrocarbon emissions flux values from individual leaks during the bagging study ranged from 0.07 g/s to 5.77 g/s. Summing all measured leaks for each individual barge yielded a barge total leak rate ranging from 1.13 to 6.24 g/s.

OTM 10 Monitoring was conducted at the Port Allen lock wall from September 24 through October 9. A total of 97 barge sets passed through the lock during the OTM 10 observation period. Six events showed significant fugitive hydrocarbon emissions as measured by OTM 10 with values ranging from 0.047g/s to 3.39 g/s AM flux rate. The equipment deployed to apply the OTM 10 approach exhibited sufficient operational robustness and detection sensitivity and was judged to be useful for mid-range distance leak detection/quantification activities where compound speciation is important. Additionally, the OTM 10 technique was able to identify and assess emission rates from a range of leak sizes as long as the prevailing wind brought the emitted plume through the vertical plane of the OTM-10 measurement configuration.

In comparing the LDEQ bagging measurements with the OTM 10 measurements (different barges), the range of AM flux values found with the OTM 10 method were generally lower than the values found using the bagging method although the maximum flux values

70


September 2009

measured are comparable (3.39 g/s with the OTM 10 method and 6.24 g/s with the bagging method). The barges selected for the bagging experiments were identified as having very significant leaks from the airborne survey so they may not represent an average case whereas the OTM 10 measurements were conducted on barges moving through the lock with no selection process and therefore represent a more typical sample cross-section. This fact could help explain the lower values observed by OTM 10.

An analysis of the infrared camera observations and ORS measurements made in the lock revealed that there were seven instances where the camera detected barge leaks, but the events were not detected by the ORS measurements. However, further analysis showed that the prevailing winds during the time of these events were parallel to the ORS measurement plane, or actually contained a southerly component (the lock was located to the north of the measurement configuration), so the barge emissions were not captured by the ORS measurement configuration.

A significant output of this project is represented in the image database which provides a comparison of PGIE images of leaks with measured leak rates which helps improve the understanding of the qualitative information provided by the infrared cameras for this source category.

Emission estimates contained in this report represent a snapshot in time. Fugitive emissions from petrochemical barges are believed to vary significantly due to ambient temperature, thermal load, product mix, load state, and equipment condition.

71


September 2009

9. References

Abichou, T., J. Clark, S. Tan, J. Chanton, G. Hater, R. Green, D. Goldsmith, M. Barlaz, and N. Swan, “Uncertainties associated with the use of OTM-10 to estimate surface emissions in landfill applications.” submitted to the Journal of the Air & Waste Management Association, 2009.Hashmonay, R.A., D.F. Natschke, K.Wagoner, D.B. Harris, E.L.Thompson, and M.G. Yost, Field evaluation of a method for estimating gaseous fluxes from area sources using open-path Fourier transform infrared, Environ. Sci. Technol., 35, 2309-2313, 2001.

Hashmonay, R.A., and M.G. Yost, Innovative approach for estimating fugitive gaseous fluxes using computed tomography and remote optical sensing techniques, J. Air Waste Manage. Assoc., 49, 966-972, 1999.

Hashmonay, R.A., M.G. Yost, D.B. Harris, and E.L. Thompson, Simulation study for gaseous fluxes from an area source using computed tomography and optical remote sensing, presented at SPIE Conference on Environmental Monitoring and Remediation Technologies, Boston, MA, Nov., 1998, in SPIE Vol. 3534, 405-410.

Thoma, E.D., R.C. Shores, E.L. Thompson, D.B. Harris, S.A. Thorneloe, R.V. Varma, R.A. Hashmonay, M.T. Modrak, D.F. Natschke, and H.A. Gamble, Open path tunable diode laser absorption spectroscopy for acquisition of fugitive emission flux data, J. Air & Waste Manage Assoc., 55, 658-668 (2005).

U.S. Environmental Protection Agency, Category IV Quality Assurance Project Plan, Development of OTM 10 for Landfill Applications-Pilot Study, U.S. EPA National Risk Management Research Laboratory, Air Pollution Prevention and Control Division, Emissions Characterization and Prevention Branch, Contract No. EP-C-05-023, Work Assignment 3-13, November, 2007b.

U.S. Environmental Protection Agency, ECPB Optical Remote Sensing Facility Manual, U.S. EPA National Risk Management Research Laboratory, Air Pollution Prevention and Control Division, Emissions Characterization and Prevention Branch, Contract No. EP-C-04-023, Work Assignment 0-33, April 2004.

U.S. Environmental Protection Agency, Final Report, Measurement of Total Site Mercury Emissions from a Chlor-alkali Plant Using Open-Path UV-DOAS; EPA/R-07/077; U.S. Environmental Protection Agency, Office of Research and Development, Work Assignment No. 2-052, July, 2007a, available at: oaspub.epa.gov/eims/eimscomm.getfile?p_download_id=469120

72


September 2009

U.S. Environmental Protection Agency, Protocol for Equipment Leaks Emission Estimating, U.S. EPA Office of Air Quality Planning and Standards, Emission Standards Division, Publication No. EPA-453/R-95-017, November, 1995.

U.S. Environmental Protection Agency. Category III Quality Assurance Project Plan, Measurement of Petroleum Barge Emissions using Optical Remote Sensing, Office of Research and Development, National Risk Management Laboratory, Emission Characterization and Prevention Branch, Research Triangle Park, NC. Work Assignment No. 4-47, August 2008.

73

Investigation of Fugitive Emissions from Petrochemical ... · PDF fileEPA/600/R-09/136...

Documents

Transcript of Investigation of Fugitive Emissions from Petrochemical ... · PDF fileEPA/600/R-09/136...